LECTURE 5 THE SURVEY: SAMPLING, QUESTIONNAIRE DISTRIBUTION AND INTERVIEWING

Notes for CIS 675 videotaped lecture: Copyright, Starr Roxanne Hiltz, 1995
1. How to select a random sample
2. Questionnaire Distribution
3. The unstructured interview and interview techniques

<TD HEIGHT=100 WIDTH=100> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld002.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld001.htm">View Graphic Version </A> </TD>

CHARACTERISTICS OF A RANDOM SAMPLE CHARACTERISTICS OF A RANDOM SAMPLE

CHARACTERISTICS OF A RANDOM SAMPLE

In a random sample, every person in the population has a DEFINITE, KNOWN probability of being included. This selection is based solely on "chance," or probability methods.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld001.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld003.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld002.htm">View Graphic Version </A> </TD>

CHARACTERISTICS OF A RANDOM SAMPLE CHARACTERISTICS OF A RANDOM SAMPLE

CHARACTERISTICS OF A RANDOM SAMPLE

With a RANDOM or PROBABILITY sample, one can make a precise ESTIMATE of how accurately the sample results are likely to correspond to those for the total population.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld002.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld004.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld003.htm">View Graphic Version </A> </TD>

PROBABILITY THEORY AND SAMPLING PROBABILITY THEORY AND SAMPLING

PROBABILITY THEORY AND SAMPLING

Suppose you have a giant bin with 25,000 red marbles and 25,000 blue marbles. You mix them thoroughly and choose 1000.
How many reds would you "expect" to get?
How many chances does each marble have of being included?

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld003.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld005.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld004.htm">View Graphic Version </A> </TD>

PROBABILITY THEORY AND SAMPLING PROBABILITY THEORY AND SAMPLING

PROBABILITY THEORY AND SAMPLING

Statisticians have done sampling experiments thousands of times, and have observed that when there are really 50% red marbles, for instance, then 95% of the time, you would draw between 474 and 536 red marbles (50% + or minus 3.5%)

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld004.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld006.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld005.htm">View Graphic Version </A> </TD>

PROBABILITY THEORY AND SAMPLING PROBABILITY THEORY AND SAMPLING

PROBABILITY THEORY AND SAMPLING

In order to obtain a "representative" sample of a larger population, random or probability methods must be used.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld005.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld007.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld006.htm">View Graphic Version </A> </TD>

THE PROCESS OF RANDOM SAMPLING THE PROCESS OF RANDOM SAMPLING

THE PROCESS OF RANDOM SAMPLING

Step one in sampling is to define the population precisely-- the group of people to whom you want to generalize your results.
Example: "NJIT students"
Then construct a "sampling frame"-- a list or data base of names, addresses, or other ID's, which is equal to the population.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld006.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld008.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld007.htm">View Graphic Version </A> </TD>

The SAMPLING FRAME The SAMPLING FRAME

The SAMPLING FRAME

is the elements from which a probability sample is selected.
ELEMENTS are usually a person, represented by a name, employee number, social security number, etc. Sometimes a household, or an organization.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld007.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld009.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld008.htm">View Graphic Version </A> </TD>

The SAMPLING FRAME The SAMPLING FRAME

The SAMPLING FRAME

The sampling frame should include all elements once and only once: No duplications or omissions. You may need to work with two or three lists of elements, and combine them, eliminate duplications, and check for any remaining omissions. The list might be on paper, on cards, in a computer data base, or other forms.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld008.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld010.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld009.htm">View Graphic Version </A> </TD>

THE PROCESS OF RANDOM SAMPLING THE PROCESS OF RANDOM SAMPLING

THE PROCESS OF RANDOM SAMPLING

Step 3: Use a method of selection which is strictly according to probability or "chance," and which does not allow any bias or influence by the selector.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld009.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld011.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld010.htm">View Graphic Version </A> </TD>

SAMPLING ERROR SAMPLING ERROR

SAMPLING ERROR

AT THE .05 LEVEL OF CONFIDENCE
Reported Number of Interviews
% 3000 1000 500 100
50 2.5 3.2 4.9 10.5
20 or 80 2.0 2.6 3.9 8.4
5 or 95 1.1 1.6 2.1 4.6

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld010.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld012.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld011.htm">View Graphic Version </A> </TD>

EXPECTED SAMPLING ERROR IS BASED ON THREE FACTORS: EXPECTED SAMPLING ERROR IS BASED ON THREE FACTORS:

EXPECTED SAMPLING ERROR IS BASED ON THREE FACTORS:

1. Sample size (the most important): Sampling error decreases with the square root of the size of the sample.
2. Variability in the population: If the population is completely homogeneous on the variable you are measuring, then a sample of one would represent it perfectly.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld011.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld013.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld012.htm">View Graphic Version </A> </TD>

EXPECTED SAMPLING ERROR EXPECTED SAMPLING ERROR

EXPECTED SAMPLING ERROR

3. The sampling method. (The chart is for "simple random" samples. Other types may have greater or less sampling error for the same size sample.)

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld012.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld014.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld013.htm">View Graphic Version </A> </TD>

EXPECTED SAMPLING ERROR EXPECTED SAMPLING ERROR

EXPECTED SAMPLING ERROR

Note that sampling error is NOT generally related to the size of the population you are sampling from. Until the sample gets very large in relation to the population, it makes no difference. A sample of 1000 from a population of 100,000 has the same estimated error as a sample of 1000 from a population of 1 million.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld013.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld015.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld014.htm">View Graphic Version </A> </TD>

SAMPLING ERROR: SAMPLING ERROR:

SAMPLING ERROR:

SAMPLING ERROR is the amount of error to be expected with a given sample size and type, at a given level of confidence.
One "Standard Error" (S)= Sq. Rt. of P x Q/ N
Where: P and Q are population percentages (e.g., 90% and 10%)

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld014.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld016.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld015.htm">View Graphic Version </A> </TD>

CONFIDENCE LEVEL and CONFIDENCE INTERVAL. CONFIDENCE LEVEL and CONFIDENCE INTERVAL.

CONFIDENCE LEVEL and CONFIDENCE INTERVAL.

These two terms define sampling error
CONFIDENCE INTERVAL is the range (+ and -) within which your sample statistic is estimated to approach the true population parameter.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld015.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld017.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld016.htm">View Graphic Version </A> </TD>

PP Presentation

CONFIDENCE LEVEL is the PROBABILITY that you are wrong--that the sampling error EXCEEDS the confidence interval given. Usually the .05 level is chosen... two standard errors out on a normal curve.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld016.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld018.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld017.htm">View Graphic Version </A> </TD>

WHY are these sampling errors true? WHY are these sampling errors true?

WHY are these sampling errors true?

The laws of probability determine that the statistics, or estimates, obtained from a series of samples of the same size and from the same population, will fall around the true value for the whole population in the shape of a normal (bell-shaped) curve.
2/3 of the sample statistics obtained will fall within plus or minus one standard error; 95% + or- two standard errors; 99% within + - three standard errors.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld017.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld019.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld018.htm">View Graphic Version </A> </TD>

RANDOM SAMPLING TECHNIQUES AND SAMPLING ERROR RANDOM SAMPLING TECHNIQUES AND SAMPLING ERROR

RANDOM SAMPLING TECHNIQUES AND SAMPLING ERROR

ALL STATISTICAL TESTS ARE RELATED TO THE IDEAS OF RANDOM SAMPLING AND THE LAWS OF PROBABILITY. IF YOU DO NOT CAREFULLY FOLLOW THE PROCEDURES TO OBTAIN A TYPE OF RANDOM SAMPLE, YOU CANNOT USE INFERENTIAL STATISTICS.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld018.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld020.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld019.htm">View Graphic Version </A> </TD>

THE SIMPLE RANDOM SAMPLE THE SIMPLE RANDOM SAMPLE

THE SIMPLE RANDOM SAMPLE

EACH ELEMENT IS INDEPENDENTLY CHOSEN, BY PROBABILITY METHODS.
1. Selection by lots.
a. Assign a number to each element.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld019.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld021.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld020.htm">View Graphic Version </A> </TD>

SIMPLE RANDOM- SELECTION BY LOT SIMPLE RANDOM- SELECTION BY LOT

SIMPLE RANDOM- SELECTION BY LOT

b. Write the numbers on identical objects which may be mixed. (ping pong balls, equal sized slips of paper, poker chips, etc.)
c. MIX the numbers very thoroughly-- much more thoroughly than might be supposed necessary. A mechanical device is necessary if there are many objects to be mixed.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld020.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld022.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld021.htm">View Graphic Version </A> </TD>

SIMPLE RANDOM- SELECTION BY LOT SIMPLE RANDOM- SELECTION BY LOT

SIMPLE RANDOM- SELECTION BY LOT

d. Draw the numbers ONE BY ONE, mixing again between draws, until the desired number of cases has been selected. The cases corresponding to the numbers drawn will then be included in the sample.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld021.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld023.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld022.htm">View Graphic Version </A> </TD>

SIMPLE RANDOM SAMPLES- BY RANDOM NUMBERS SIMPLE RANDOM SAMPLES- BY RANDOM NUMBERS

SIMPLE RANDOM SAMPLES- BY RANDOM NUMBERS

Procedure 2- SELECTION BY A TABLE OF RANDOM NUMBERS. (either through human use or computer use of a random number generator).

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld022.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld024.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld023.htm">View Graphic Version </A> </TD>

SIMPLE RANDOM SAMPLES- BY RANDOM NUMBERS SIMPLE RANDOM SAMPLES- BY RANDOM NUMBERS

SIMPLE RANDOM SAMPLES- BY RANDOM NUMBERS

A table of random numbers is simply a list of digits generated by some mechanical mixing and drawing procedure, so that each digit has an equal chance of coming up on each draw. Statistics and Methods books put some pages of random numbers in the appendix.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld023.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld025.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld024.htm">View Graphic Version </A> </TD>

A PARTIAL PAGE OF RANDOM DIGITS A PARTIAL PAGE OF RANDOM DIGITS

A PARTIAL PAGE OF RANDOM DIGITS

88 83 55 44 86 23 76 80 61 56
04 11 10 84 08 05 46 26 92 00
90 82 29 70 22 17 71 90 42 07
56 19 68 00 91 82 06 76 34 00
07 47 74 46 06 17 98 54 89 11

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld024.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld026.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld025.htm">View Graphic Version </A> </TD>

SYSTEMATIC RANDOM SAMPLE: SELECTION AT INTERVALS ON A LIST. SYSTEMATIC RANDOM SAMPLE: SELECTION AT INTERVALS ON A LIST.

SYSTEMATIC RANDOM SAMPLE: SELECTION AT INTERVALS ON A LIST.

First examine the list for some sort of ordering bias or trend... Cannot be used if there is some sort of biasing order that may correspond to the desired sampling interval.
1 Number the list
2 Determine total sample size desired.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld025.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld027.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld026.htm">View Graphic Version </A> </TD>

3. Determine the SAMPLING INTERVAL, called "K" 3. Determine the SAMPLING INTERVAL, called "K"

3. Determine the SAMPLING INTERVAL, called "K"

N of cases on list

K= _________________

Sample size

Example:

K= 500 names = 4

_________

125 sample

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld026.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld028.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld027.htm">View Graphic Version </A> </TD>

SYSTEMATIC RANDOM SAMPLE: SELECTION AT INTERVALS ON A LIST. SYSTEMATIC RANDOM SAMPLE: SELECTION AT INTERVALS ON A LIST.

SYSTEMATIC RANDOM SAMPLE: SELECTION AT INTERVALS ON A LIST.

4. Make a random choice of starting point within the first K.
5. Continue selecting every Kth after that. When bottom of a page is reached, or end of a file drawer, continue the count at the beginning of the next.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld027.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld029.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld028.htm">View Graphic Version </A> </TD>

Systematic Sampling Systematic Sampling

Systematic Sampling

Systematic sampling can be done in "inches" from cards in file drawers, rather than names on a list.
Example: sample of 3000 policy holders from file drawers full of 2.5 million, one per card.
Compress cards
Count cards per foot and total number of feet to get N of cases
Select one card every K inches.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld028.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld030.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld029.htm">View Graphic Version </A> </TD>

MULTI-STAGE RANDOM SAMPLES MULTI-STAGE RANDOM SAMPLES

MULTI-STAGE RANDOM SAMPLES

Example 1: A two-stage area sample
1. Define the area to be included. Using an up-to-date map, divide the entire area into mutually exclusive smaller areas or block. Number them (usually done serially and i serpentine fashion).
2. Select the sample blocks, with simple random or systematic random design.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld029.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld031.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld030.htm">View Graphic Version </A> </TD>

TWO STAGE AREA SAMPLE TWO STAGE AREA SAMPLE

TWO STAGE AREA SAMPLE

3. List each dwelling in the sampled blocks. This usually requires fieldwork.
4. Select the sample dwellings by a random method. Then establish rule for which individual within each sampled dwelling will actually be interviewed.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld030.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld032.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld031.htm">View Graphic Version </A> </TD>

OTHER EXAMPLES of MULTI-STAGE SAMPLES: OTHER EXAMPLES of MULTI-STAGE SAMPLES:

OTHER EXAMPLES of MULTI-STAGE SAMPLES:

Multi-stage national sample (states; counties; townships; blocks. Selection of households on sampled blocks is then usually done proportionally to population, as determined by census figures.
Corporation: e.g., IBM
Step 1: country; Step 2 office locations; Step 3 divisions; Step 4 Staff groups...

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld031.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld033.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld032.htm">View Graphic Version </A> </TD>

Advantages of multi-stage sampling: Advantages of multi-stage sampling:

Advantages of multi-stage sampling:

Sampling lists and numbering required only for members of selected sampling units.
If sampling units are geographically defined, cuts down field costs for travel for personal interviews.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld032.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld034.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld033.htm">View Graphic Version </A> </TD>

DISADVANTAGE OF MULTI- STAGE SAMPLING: DISADVANTAGE OF MULTI- STAGE SAMPLING:

DISADVANTAGE OF MULTI- STAGE SAMPLING:

Larger sampling error than for one-stage samples of same size.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld033.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld035.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld034.htm">View Graphic Version </A> </TD>

OTHER RANDOM SAMPLING METHODS OTHER RANDOM SAMPLING METHODS

OTHER RANDOM SAMPLING METHODS

STRATIFIED
Proportionate
Disproportionate- with weights
CLUSTER

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld034.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld036.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld035.htm">View Graphic Version </A> </TD>

OPTIONS FOR QUESTIONNAIRE DISTRIBUTION OPTIONS FOR QUESTIONNAIRE DISTRIBUTION

OPTIONS FOR QUESTIONNAIRE DISTRIBUTION

CRITERION: GROUP COMPLETION INDIVIDUAL INTERVIEW
Response Rate highest (95% +) High (80%)
Possible Limited to existing groups Unlimited populations groups
Understanding Low High by low SES
Depth and detail Low High in answers
Time and cost Medium Most expensive
Truth on touchy issues High Medium

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld035.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld037.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld036.htm">View Graphic Version </A> </TD>

OPTIONS FOR QUESTIONNAIRE DISTRIBUTION OPTIONS FOR QUESTIONNAIRE DISTRIBUTION

OPTIONS FOR QUESTIONNAIRE DISTRIBUTION

CRITERION MAILED QUESTIONNAIRE TELEPHONE (CATI)
Response Rate LOW (10%-50%) Medium (20%- 70%)
Possible Slightly limited Slightly limited populations
Understanding by Very low Medium low ses
Depth and detail Low (to high-- Delphi) Medium in answers
Time and cost Medium Expensive
Truth on touchy Medium Low

issues

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld036.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld038.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld037.htm">View Graphic Version </A> </TD>

COMPUTER (self-administered) COMPUTER (self-administered)

COMPUTER (self-administered)

Response Rate ?? (?medium to high?)
Possible populations PC owners
Understanding by low None

SES

Depth and detail in ? (high?) answers
Time and cost Medium
Truth High (?highest?)

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld037.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld039.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld038.htm">View Graphic Version </A> </TD>

INTERVIEWER ATTITUDES TOWARD COMPUTER-ASSISTED PERSONAL INTERVIEWING (CAPI) INTERVIEWER ATTITUDES TOWARD COMPUTER-ASSISTED PERSONAL INTERVIEWING (CAPI)

INTERVIEWER ATTITUDES TOWARD COMPUTER-ASSISTED PERSONAL INTERVIEWING (CAPI)

Couper & Burt, Social Science Computer Review, Spring 1994.
Provides example of a survey-based study AND
Data on Computer-Assisted Interviewing

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld038.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld040.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld039.htm">View Graphic Version </A> </TD>

CAPI and CATI CAPI and CATI

CAPI and CATI

CATI is computer-assisted Telephone interviewing; it is a form of CAPI that adds features for telephone call management to the software, but otherwise is similar
At its simplest, CAPI involves the administration of a questionnaire by an interviewer using a laptop computer that is usually taken into the respondent's home or office.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld039.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld041.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld040.htm">View Graphic Version </A> </TD>

RESEARCH QUESTIONS: RESEARCH QUESTIONS:

RESEARCH QUESTIONS:

Older women are more likely to express "computer anxiety"
Older women are the most frequent interviewer labor force; they like the opportunity to get out and interact with other people
Does computer anxiety among these interviewers create problems with CAPI? How does this change with experience?

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld040.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld042.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld041.htm">View Graphic Version </A> </TD>

SAMPLE SAMPLE

SAMPLE

380 out of sample of 404 interviewers who were trained for CAPI from those participating in first use of CAPI by BLS (Bureau of Labor Statistics) ) Note: "strategic research site."

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld041.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld043.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld042.htm">View Graphic Version </A> </TD>

SAMPLE SAMPLE

SAMPLE

Note discussion of sample limitations: though instructed to select normal mix of interviewers for CAPI training, offices 'may have assigned their better interviewers to those who were more positively disposed toward computers."

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld042.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld044.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld043.htm">View Graphic Version </A> </TD>

STUDY DESIGN (LONGITUDINAL) STUDY DESIGN (LONGITUDINAL)

STUDY DESIGN (LONGITUDINAL)

Q 1 sent with self-study material and laptop computer, before they were to report for training
Q2 mailed after completion of two assignments, about 2 months later

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld043.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld045.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld044.htm">View Graphic Version </A> </TD>

SCALES SCALES

SCALES

21 LIKERT-TYPE (FIVE POINT) ITEMS, E.G.
I am nervous about taking a computer into certain areas for fear of its being stolen [R] (R means scoring was reversed so that disagreement with a negative item was given a high score)

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld044.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld046.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld045.htm">View Graphic Version </A> </TD>

RESULTS RESULTS

RESULTS

Follows usual pattern of reporting univariate, bivariate, and multivariate analyses
Overall pre-training attitudes positive; mean of 3.76. Mean T2 results even more positive

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld045.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld047.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld046.htm">View Graphic Version </A> </TD>

RESULTS RESULTS

RESULTS

T1 Means by interviewer characteristics show the more experienced the interviewer, the less positive the expectation;
Significant bivariate differences between males and females at T1 disappear in multiple regression, where only education and computer experience remain significant.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld046.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld048.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld047.htm">View Graphic Version </A> </TD>

RESULTS RESULTS

RESULTS

Some changes between T1 and T2 indicate improvements to system that must be made, e.g. from 24% to 39% agreeing that "Using a computer slows me down during interviews"

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld047.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld049.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld048.htm">View Graphic Version </A> </TD>

THE UNSTRUCTURED OR FOCUSED INTERVIEW: A QUALITATIVE METHOD THE UNSTRUCTURED OR FOCUSED INTERVIEW: A QUALITATIVE METHOD

THE UNSTRUCTURED OR FOCUSED INTERVIEW: A QUALITATIVE METHOD

Uses open-ended questions to explore a topic or a few related topics of interest. The exact questions used, their wording and order, will vary from person to person, in order to draw out of each interviewee as much as they know about the subject of interest.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld048.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld050.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld049.htm">View Graphic Version </A> </TD>

SOME EXAMPLES OF USES OF UNSTRUCTURED INTERVIEWS: SOME EXAMPLES OF USES OF UNSTRUCTURED INTERVIEWS:

SOME EXAMPLES OF USES OF UNSTRUCTURED INTERVIEWS:

1. Providing background for defining a conceptual model and stating an objective in the early exploratory stages of a study.-- e.g., understanding organizational context, existing system, functional specifications, to begin to define new system requirements.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld049.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld051.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld050.htm">View Graphic Version </A> </TD>

PP Presentation

2. Follow-up on a protocol analysis
3. Serving as a basis for constructing more structured questionnaires by showing the general framework within which respondents think and the words they use.
4. Explaining ambiguous or incomplete findings from a large survey or experiment.
5. Debriefing subjects in a controlled experiment.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld050.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld052.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld051.htm">View Graphic Version </A> </TD>

PROCEDURES FOR CONDUCTING UNSTRUCTURED INTERVIEWS PROCEDURES FOR CONDUCTING UNSTRUCTURED INTERVIEWS

PROCEDURES FOR CONDUCTING UNSTRUCTURED INTERVIEWS

1. An INTERVIEW GUIDE is prepared which lists the questions and subject areas that will be covered. However, the amount of time spent on each question, and the order followed, vary.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld051.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld053.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld052.htm">View Graphic Version </A> </TD>

PROCEDURES PROCEDURES

PROCEDURES

The interview has a CONVERSATIONAL structure, and the PROBE is used for those questions about which respondents seem particularly informed or responsive.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld052.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld054.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld053.htm">View Graphic Version </A> </TD>

PROCEDURES FOR CONDUCTING UNSTRUCTURED INTERVIEWS PROCEDURES FOR CONDUCTING UNSTRUCTURED INTERVIEWS

PROCEDURES FOR CONDUCTING UNSTRUCTURED INTERVIEWS

Questions evoking minimal response are sometimes restated in different words. In short, these interviews are customized to each respondent and interviewing situation to try to elicit all experiences and attitudes relevant to the topic being investigated.
Because the subject is encouraged to talk freely about the topic, these interviews usually last one to two hours, and may even be continued over two or more sessions.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld053.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld055.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld054.htm">View Graphic Version </A> </TD>

2. CONDUCTING THE INTERVIEW 2. CONDUCTING THE INTERVIEW

2. CONDUCTING THE INTERVIEW

The conversational structure and use of probes differ from in-person interviewing using a totally structured or fixed questionnaire. However, general principles of inteview techniques are similar.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld054.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld056.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld055.htm">View Graphic Version </A> </TD>

2. CONDUCTING THE INTERVIEW: 2. CONDUCTING THE INTERVIEW:

2. CONDUCTING THE INTERVIEW:

You need to
a. Establish LEGITIMACY and evoke the interest of the subject with your introduction and first few questions, which should be interesting but non-threatening (GET THE INTERVIEW).

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld055.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld057.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld056.htm">View Graphic Version </A> </TD>

2. CONDUCTING THE INTERVIEW: 2. CONDUCTING THE INTERVIEW:

2. CONDUCTING THE INTERVIEW:

b. Maintain CONTROL-- without being abrasive or authoritative, you must guide the interviewee into productive directions-- do not allow evasions, irrelevancy, or superficial answers. The PROBE is the chief implement here.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld056.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld058.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld057.htm">View Graphic Version </A> </TD>

2. CONDUCTING THE INTERVIEW: 2. CONDUCTING THE INTERVIEW:

2. CONDUCTING THE INTERVIEW:

c. Build RAPPORT- by being non-judgmental, supportive, giving verbal rewards.
(Comments like "I see... That's interesting...)
d. Obtain an accurate but unobtrusive record of the responses.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld057.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld059.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld058.htm">View Graphic Version </A> </TD>

OBTAINING THE INTERVIEW OBTAINING THE INTERVIEW

OBTAINING THE INTERVIEW

Dress appropriately. Carry an ID. Offer to set up an appointment whenever it is convenient. Stress the importance of the study and use advance publicity or contacts if possible.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld058.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld060.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld059.htm">View Graphic Version </A> </TD>

The INTRODUCTION or PITCH: Explain: The INTRODUCTION or PITCH: Explain:

The INTRODUCTION or PITCH: Explain:

WHO- give your name
WHAT you are studying
WHY- for what sponsor or purpose
HOW LONG it will take
Assure CONFIDENTIALITY
Then: assume cooperation and ask the first question

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld059.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld061.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld060.htm">View Graphic Version </A> </TD>

"VOLUNTARY" PARTICIPATION "VOLUNTARY" PARTICIPATION

"VOLUNTARY" PARTICIPATION

If we define a "volunteer" as "anyone who comes to the interviewer and asks to be included," representative studies could not be done.
If we mean that anyone who is "busy" or "doesn't know anything about that" on the first contact should not be interviewed, then representative studies would also be impossible.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld060.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld062.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld061.htm">View Graphic Version </A> </TD>

PP Presentation

PROMISES- to give them a copy of the report, or to help them out in some way- are OK AS LONG AS YOU ARE PREPARED TO CONSCIENTIOUSLY FOLLOW THROUGH.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld061.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld063.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld062.htm">View Graphic Version </A> </TD>

"VOLUNTARY" PARTICIPATION "VOLUNTARY" PARTICIPATION

"VOLUNTARY" PARTICIPATION

Ethically, most "volunteers" are actually "persuaded cooperators" or at least nonresisters to gentle coercion. There are many legitimate claims that can be made to persuade the respondent to be interviewed, such as an appeal to the respondent's altruism, the chance to have one's preferences recorded and counted, the interesting nature of the topic or experience. Outright coercion is not ethical, but gentle persuasion is OK.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld062.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld064.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld063.htm">View Graphic Version </A> </TD>

CONDUCTING THE INTERVIEW CONDUCTING THE INTERVIEW

CONDUCTING THE INTERVIEW

Basic rules:
THE SITUATION: Find a place for the interview (or protocol analysis) which provides as much privacy as possible. It must be quiet, and nobody else should be watching or listening.
NEUTRALITY- interviewer must behave in such a way as not to affect responses. Questions must not be biased or leading.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld063.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld065.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld064.htm">View Graphic Version </A> </TD>

CONDUCTING THE INTERVIEW CONDUCTING THE INTERVIEW

CONDUCTING THE INTERVIEW

FAMILIARITY with the interview guide-- You should know the question well enough before beginning so that you can refer to the guide (or questionnaire) for cues, but SPEAK TO the respondent most of the time, not read questions. PRACTICE FIRST.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld064.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld066.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld065.htm">View Graphic Version </A> </TD>

USING AN INTERVIEW GUIDE USING AN INTERVIEW GUIDE

USING AN INTERVIEW GUIDE

Try to structure the flow of questions so that it is as much like a natural conversation as possible. Use transitional phrases when changing topic. Refer back to respondent's remarks to show that you have been listening.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld065.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld067.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld066.htm">View Graphic Version </A> </TD>

Put on your "WAGS" Put on your "WAGS"

Put on your "WAGS"

Warmth
Acceptance
Genuineness
Support

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld066.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld068.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld067.htm">View Graphic Version </A> </TD>

PP Presentation

WARMTH: Establish rapport. Show a positive interest in the subject. Be pleasant, receptive, patient, and don't interrupt. SMILE.
Reassure respondent that you're interested in and respect what he's saying.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld067.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld069.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld068.htm">View Graphic Version </A> </TD>

ACCEPTANCE ACCEPTANCE

ACCEPTANCE

This means taking the person as he is and where he is. It means understanding that the way he is and the ways he acts are the product of all he has known and that they are rarely consciously intentioned.
Reassure the respondent that inability to answer certain questions is no loss of prestige. (Suggest that perhaps you and many others can't answer them either.)

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld068.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld070.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld069.htm">View Graphic Version </A> </TD>

ACCEPTANCE ACCEPTANCE

ACCEPTANCE

Part of Acceptance is EMPATHY- feeling WITH another person, being able to "be in his shoes." Show caring and concern... caring is love in its giving, protective, nurturant aspects. In short... Act like a therapist!

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld069.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld071.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld070.htm">View Graphic Version </A> </TD>

GENUINENESS GENUINENESS

GENUINENESS

To have a sense of who you are and what you believe in.. Do not allow empathy to lead you to lie to try to increase rapport.
e.g.- If asked what you think, say you are interested in what the RESPONDENT thinks, and if he or she really wants to know what you think, ask again when the interview is completed.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld070.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld072.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld071.htm">View Graphic Version </A> </TD>

SUPPORT SUPPORT

SUPPORT

Don't push the respondent or threaten his ego. Help him to put things in his own words, never yours. Give him time to explain something further if he's uneasy about what he's already said.
Never indicate whether you agree or disagree with what the respondent says... all you want to do is UNDERSTAND it.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld071.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld073.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld072.htm">View Graphic Version </A> </TD>

CONTROLLING THE INTERVIEW CONTROLLING THE INTERVIEW

CONTROLLING THE INTERVIEW

Don't let the interviewee get off the subject. Guide him or her toward giving a full and consistent report.
This means PROBING.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld072.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld074.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld073.htm">View Graphic Version </A> </TD>

Some types of non-directive probes: Some types of non-directive probes:

Some types of non-directive probes:

1. BRIEF, NEUTRAL EXPRESSIONS OF UNDERSTANDING AND INTEREST
Examples: I see; um-hmm; yes; I understand; nodding of head.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld073.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld075.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld074.htm">View Graphic Version </A> </TD>

PROBES PROBES

PROBES

2. Brief expectant pauses
Short pauses (2-3 seconds) will usually encourage respondent to go on without your having to say anything. However, long pauses (in excess of 10-15 seconds) make the respondent feel uncomfortable.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld074.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld076.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld075.htm">View Graphic Version </A> </TD>

PROBES PROBES

PROBES

3. NEUTRAL REQUESTS FOR MORE INFORMATION
Examples: How do you mean? I'd like to know more of your thinking on that. What do you have in mind there? Is there anything else? Can you tell me more about that?

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld075.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld077.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld076.htm">View Graphic Version </A> </TD>

4. ECHO OR NEAR REPETITION OF THE RESPONDENT'S WORDS 4. ECHO OR NEAR REPETITION OF THE RESPONDENT'S WORDS

4. ECHO OR NEAR REPETITION OF THE RESPONDENT'S WORDS

Example: Respondent: "I look at this screen and it makes my eyes hurt."
Interviewer: "It makes your eyes hurt?"

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld076.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld078.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld077.htm">View Graphic Version </A> </TD>

5. SUMMARIZING OR REFLECTING RESPONDENT EXPRESSIONS 5. SUMMARIZING OR REFLECTING RESPONDENT EXPRESSIONS

5. SUMMARIZING OR REFLECTING RESPONDENT EXPRESSIONS

Examples would follow respondent statements, stating the interviewer's understanding of a key feeling or meaning. Such summaries often begin with phrases like, "You feel that... or "You mean that..."

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld077.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld079.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld078.htm">View Graphic Version </A> </TD>

PP Presentation

CAUTION: When accurate summaries are given within a framework of acceptance, they may stimulate further conversation. However, if the interviewer inaccurately summarizes or shows non permissiveness, the respondent may well become defensive.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld078.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld080.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld079.htm">View Graphic Version </A> </TD>

6. REQUESTS FOR SPECIFIC KINDS OF ADDITIONAL INFORMATION 6. REQUESTS FOR SPECIFIC KINDS OF ADDITIONAL INFORMATION

6. REQUESTS FOR SPECIFIC KINDS OF ADDITIONAL INFORMATION

Examples: WHY do you think that is so?
HOW did that become clear to you?
WHEN was that?

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld079.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld081.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld080.htm">View Graphic Version </A> </TD>

7. REQUESTS FOR CLARIFICATION 7. REQUESTS FOR CLARIFICATION

7. REQUESTS FOR CLARIFICATION

Examples: I'm not clear on that. Could you explain what you mean?
Could you give me some examples?
Often, the interviewer may note inconsistencies between responses and wish to clarify them. Usually, a permissive attitude, summary of the alleged inconsistencies, and a request for clarification are all that are necessary. For example:

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld080.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld082.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld081.htm">View Graphic Version </A> </TD>

REQUESTS FOR CLARIFICATION REQUESTS FOR CLARIFICATION

REQUESTS FOR CLARIFICATION

Interviewer: Could we go back over something for a moment? Earlier you said that your husband worked for a research firm, but now you say that he is a teacher. Could you please clarify this for me?

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld081.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld083.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld082.htm">View Graphic Version </A> </TD>

8. REPETITION OF A PRIMARY QUESTION 8. REPETITION OF A PRIMARY QUESTION

8. REPETITION OF A PRIMARY QUESTION

Examples
Interviewer: What kind of work do you do?
Respondent: I work at NJIT.
Interviewer: I see. What kind of work do you do there?

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld082.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld084.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld083.htm">View Graphic Version </A> </TD>

RECORDING THE INTERVIEW RECORDING THE INTERVIEW

RECORDING THE INTERVIEW

Take a clipboard, notepad, TWO pens or pencils, tape recorder, extra tapes and batteries.
Record a full, verbatim interview.
Hand notes might be used to note what actions or facial expressions are being used; notes to yourself to follow up later.
Explain that the tape will be erased after the interview is completed, and that it is being used only to make sure that you accurately record the respondent's words.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld083.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> <A HREF = "tsld085.htm">Next slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld084.htm">View Graphic Version </A> </TD>

RECORDING THE INTERVIEW RECORDING THE INTERVIEW

RECORDING THE INTERVIEW

The interviewer must transcribe the interview.
Do this as soon after the interview as possible, so that you can remember parts that may be garbled in the tape.
Double space and leave big margins, so that you can add interpretations and classifications.

<TD HEIGHT=100 WIDTH=100> <A HREF = "tsld084.htm">Previous slide </A> </TD> <TD HEIGHT=100 WIDTH=100> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "index.htm">Back to the first slide </A> </TD> <TD HEIGHT=100 WIDTH=150> <A HREF = "sld085.htm">View Graphic Version </A> </TD>