the surrey business school teaching how to sample in research: some things that might help mark nk...
TRANSCRIPT
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Teaching how to sample in research:
some things that might help
Mark NK Saunders
The Surrey Business Schoolwww.surrey.ac.uk/sbs
A sampling issue?
In February 2011 the UK’s Outdoor Media Centre launched its ‘Hall of Fame’ competition to identify the 100 best advertising posters of all time. Working with the History of Advertising Trust, they generated a list of 500 posters. This was reduced to a shortlist of 228 posters by a committee of media and creative experts together with the editor of weekly magazine - Campaign. These were displayed on a dedicated website www.outdoorhalloffame.co.uk. Creative agencies, media planners, advertisers, media owners and the general public were invited in an article in Campaign to go to the web site, view the advertising campaigns and cast their votes for what they considered to be the best outdoor posters. Each person was able to cast a total of ten votes, the best advertisements being “chosen after more than 10,000 reader votes”.
The Surrey Business Schoolwww.surrey.ac.uk/sbs
How might sample selection have affected the results?
12
3
4
5
67
8
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Session PlanSampling is not boring –attracting attentionSampling does not always involve balls and urns –introducing non probability samplingSampling does not always require a sampling frameSampling and statistics –misconceptions about probability samplingThe sample selected matters –the link with research question The how many question -the issue of sample size
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Attracting attention
“A small test of 50 women found
almost nine out of 10 of them said their skin looked more luminous after using it, while 72 per cent said fine lines appeared less visible.”
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Introducing non probability sampling (1)
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Introducing non probability sampling (2)
Some implied research questions (students’ suggestions)
How do people manage their careers?
What is the best way to manage a career?
Guildford
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Sampling does not always require a sampling frame…
Data collected by… Catching commuter train from
Guildford to London 2nd class ticket 5 weeks – Monday to Friday Interviewing 5 people on average
each journey
Unclear regarding… Time of year How selected participants
The Surrey Business Schoolwww.surrey.ac.uk/sbs
The sample selected matters
Characteristics of population from which sample selected
Catching commuter train (commuters?)
Living near Portsmouth – Guildford – London line
Working in London
Working “9 to 5”
What else?
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Sample selected matters (2)…Generalisabiity: the way in which the sample is drawn affects the extent
to which the findings can be applied (generalised) or other settings
Sample from: Commuters Living south of London Unlikely to be senior
managers Working “9 to 5” Unlikely to be employed
in manufacturing
Research Question People Living anywhere With a career Doing any type of job
The Surrey Business Schoolwww.surrey.ac.uk/sbs
So what have we done so far?
Hopefully sampling is not boring –real issue in research
Sampling does not always involve balls and urns –introduced non probability sampling and shown in relation to this…
…sampling does not always rely on a sampling frame –can’t get a list of passengers easily
The sample selected matters…
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Testing understanding…What is the relationship between a population and a sample?
1 2 3 4
0% 0%0%0%
1. The sample is a sub set of the population
2. The sample is the population
3. The sample is always representative of the population
4. The sample is the method by which you select the population
15
www.rwpoll.com
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Samples … realities in management research (& the Guildford study)
REALITIES
Don’t always have sampling frame
Sampling frame often restricted
Difficult to show sample truly random
CONSEQUENCES
Inferences (including statistical) confined to actual population from which sample drawn –can’t generalise statistically
May not be able to answer certain questions … rephrase the question!
Population assumed itself to be a random sample from larger population or
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Unclear how non-probability sample was drawn
Sampling
Probability Non-probability
Volunteer Haphazard
Snowball Selfselection Convenience
Heterogeneouspurposive
Homogeneouspurposive
Critical casepurposive
Typicalcase
purposiveTheoretical
Purposive
Extremecase
purposive
Quota
Quota
But different non-
probability techniques would have
differing implications
The Surrey Business Schoolwww.surrey.ac.uk/sbs
The how many question –what non-probability sample size?
REMEMBER –making generalisations to theory not about a population (except with quota samples)
Sample size depends upon research question and objectives… what need to find out what will be useful what is credible what can be done within available resources
Continue until reach data saturation
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Non-probability sample size ...more detail
Author Nature of study Size
Bertaux (1981) Qualitative 15
Kvale and Brinkmann (2009) Interviews 5-25
Bernard (2000); Morse (1994) Ethnographic 35 - 36
Creswell (1998); Morse (1994) Grounded theory 20 - 35
Creswell (1998); Morse (1994) Phenomenological 5 - 25
Guest et al. (2006); Kuzel (1992); Romney, Batchelder and Weller (1986)
Considering a homogenous population
4 – 12
Kuzel (1992); Creswell (2007) Considering a heterogeneous population
12 - 30
Source: Saunders (2012)
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Non probability sample selection
Justifying choices Following a recipe
for Quota sampling
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Misconceptions about probability sampling –simple random sampling
Selecting a sample using simple random sampling will select the sample evenly from the population
If a truly random process is used to select a sample from a population, then the resulting sample will be just like the population, but smaller –a miniature replica of the population.
The size of the sample should be proportional to the size of the population. -the ratio of sample size to population size needs to be considered when deciding how large the sample should be.
The larger the sample the greater the accuracy
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Misconceptions…-simple random sampling selects evenly
Last five UK lottery draws
5 17 32 39 42 40
9 22 31 34 35 43
7 10 12 15 25 42
2 9 27 33 39 40
8 14 15 17 20 47
Look at the pattern of numbers
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Misconceptions about probability sampling –miniature replica
Write down your student number.
Then answer the questions on the slide for the last digit of your student number115678 = 8113459 = 9 etc etc
Then, I will show the questions again but this time we will ALL answer them
In comparing the answers, we will be able to see realities of random sampling…. (and of questionnaires!)
(Idea from Prof Sam Warren)
The Surrey Business Schoolwww.surrey.ac.uk/sbs
I often receive bad service when I am shopping in England (student number ending 1, 4, 5)
1 2 3 4 5
20%
17%
20%
17%
27%
1. Strongly agree
2. Agree
3. Neither agree or disagree
4. Disagree
5. Strongly disagree
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Results –a miniature replica?
Students Number Strongly agree
Agree Neither Disagree Strongly disagree
1,4,5 57 20% 17% 27% 17% 20%2,3,8 63 12% 19% 29% 23% 17%6,7,0 47 19% 18% 26% 19% 18%All (incl. 9)
162 17% 18% 28% 20% 17%
Impact of larger
sample size
Variation in number
responding
Variations in response
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Misconception: Sample size is proportional to population(1)
Sample size for 95% certainty with 5% margin of error
0
50
100
150
200
250
300
350
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Total Population
Sam
ple
siz
e
The problem of making statistical inferences from a small probability sample
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Probability sample and population size (2)
0
50
100
150
200
250
300
350
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Total Population
Sam
ple
siz
e
population 5% 2%
10,000 370 1936
1,000,000 384 2395
10,000,000 384 2400
Sample size for 95% certainty with 5% margin of error
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Misconception: larger the sample size greater the accuracy
0
0.01
0.02
0.03
0.04
0.05
0 2000 4000 6000 8000 10000
The larger the probability sample size, the more confident can be that represents the true population for a given population size.
population 5% 2%
100 79 96
200 132 185
500 217 226
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Different probability sampling techniques would have different implications (given knew total population)
Systematic Stratifiedrandom
ClusterSimplerandom
Multi-stage
Sampling
Probability
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Probability sampling –following a recipe
How to do it most text books -just follow the recipe…-use practice exercise-justify choices
Key issue is understanding why and implications of what do-discuss case studies...-use multiple choice questions to check http://wps.pearsoned.co.uk/ema_uk_he_saunders_resmethbus_5/111/28550/7309036.cw/index.html
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Probability samples … statistical inference Inferential statistics allow
general statements to be made about the population form the sample providing data normally distributed
Betting about the probability of being wrong or inferring invalid conclusions
“It is a sad fact that if one knows nothing about the probability of occurrence for aparticular sample of units for observation, very little of the inferential machinery … can be applied”
(Hays 1994: 227)
The Surrey Business Schoolwww.surrey.ac.uk/sbs
But beware of the effect of sample size on practical significance…
The larger the sample size the more likely your results are statistically significant
Statistically significant result - unlikely to have occurred by chance
Practically significant result – practically meaningful in real world
Jacob Cohen
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Response ratesMost important aspect is sample is representative of ...
High response rate can help this
Non-responses different from rest of population as did not respond!
Non response due to 4 interrelated problems:
refusal to respond
ineligibility to respond
inability to locate respondent
respondent located but unable to make contact
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Bibliography (1)
American Association for Public Opinion Research (2008) Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys (5th edition) Lenexa, KA: AAPOR
Baruch Y and Holtom BC (2008) Survey response rate levels and trends in organizational research Human Relations 6.8 1139-60.
Groves, R.M., Dillman, D.A., Eltinge, J.L. and Little, R.J.A. (eds.) (2001) Survey Non-response. New York: John Wiley.
Guest, G., Bunce, A. and Johnson, L. (2006). How many interviews are enough? An experiment with data saturation and validity. Field Methods. Vol. 18, No. 1, pp. 59-82
Lance, C.E. & Vandenberg, R. J. (2009), Statistical and Methodological Myths and Urban Legends. Routledge, London:Taylor & Francis.
The Surrey Business Schoolwww.surrey.ac.uk/sbs
Bibliography (2)
Meyer, J.H.F., Shanahan, M.P. & Laugksch, R.C. (2005), “Students’ Conceptions of Research. I: A qualitative and quantitative analysis”, Scandinavian Journal of Educational Research, 49. 3, 225-244.
Patton, M.Q. (2002) Qualitative Research and Evaluation Methods. 3rd edn. Thousand Oaks, CA: Sage, Ch.5
Rogelberg SG and Stanton JM (2007) Introduction: Understanding and dealing with organizational survey non-response Organizational Research Methods 10.2 195-209
Saunders, M.N.K., Lewis, P. and Thornhill, A. (2012) Research Methods for Business Students. 6th edn. London: FT Prentice Hall, Ch.7.
Saunders M.N.K. (2012) ‘Choosing research participants’ in G Symons and C Cassell (eds) The Practice of Qualitative Organizational Research: Core Methods and Current Challenges.London: Sage 37-55