Science Buddies: "Ask an Expert"

Eek! I don't know anything at all about statistics..

I have a project where I am looking at the perception of closeness in adolescent romantic relationships, based on gender, age, and length of relationship. The survey I distributed can be found on http://maxpages.com/loveadoles/Survey

How am I supposed to analyze my data? What statistical measure should I use?


The program I am working with is Excel.. If you need to see my data for a better understanding of what I'm talking about, I'll email it to you

thank you!

adrienne wrote:Eek! I don't know anything at all about statistics..

I have a project where I am looking at the perception of closeness in adolescent romantic relationships, based on gender, age, and length of relationship. The survey I distributed can be found on http://maxpages.com/loveadoles/Survey

How am I supposed to analyze my data? What statistical measure should I use?


The program I am working with is Excel.. If you need to see my data for a better understanding of what I'm talking about, I'll email it to you

thank you!

Hi adrienne!
What statistical test you use really depends on your hypothesis...and is your data numerical or in another layout? it might be easier to assign values for closeness, for example, on a scale from 1-10. Based on your hypothesis (depending if you think that the expected average closeness value is 1,2,3,4, etc) then you can figure out the observed average...you can then do a number of tests with a sampling distribution to figure out the percentile and to see if your hypothesis was right...wikipedia also has an article that lists many of the statistical methods for hypothesis testing-
http://en.wikipedia.org/wiki/Statistica ... is_testing
hope this helped, good luck!

I assume since you said in the posting subject you said this was a project for an Intel fair, you are trying to get into Intel ISEF. If this is incorrect, please let me know.

The first statistical concern that comes to mind when I see your project is that it is based on a survey administered over the internet. Surveys given on the internet are often flawed - not all of them are, but there are some definite issues that need to be taken account when doing a survey over the internet. These issues largely arise because you can never be sure who is actually responding to the survey; in addition, you have a self-selecting sample population. That is, you don't decide who takes your survey - people decide to take your survey. In many cases in which data is collected using an internet-based survey posted on a random webpage, the "results" are quite a bit out of line with the population. Big surveys, like the Harris Poll, spend lots and lots of time and money figuring out how to interpret and use the data they get from internet surveys.

That being said, there are ways to do meaningful surveys on the internet. I would recommend doing research about this subject.

To answer your question about how to analyze your data, after looking at your survey, I see that you have 1 categorical variable and 4 quantitative variables. So, I'm guessing that you want to look for correlations between the picture that describes the relationship and the four other variables. The first thing that comes to mind is a two-way table to examine the relationship between the categorical variable and perceived closeness and a scatterplot to examine the relationship between the quantitative variables and perceived closeness.

Excel has the capabilities to create both the tables and the scatterplots. Once you have done some exploratory data analysis on the data, we can start to talk about doing some inference.

If you need help with the statistical aspects of your project, or don't understand what I've posted, please let me know. I would be more than happy to help you.

Hi Adrienne,
For some additional opinions about the pros and cons of online surveys, please visit this thread:
http://www.sciencebuddies.com/mentoring ... php?t=2666
There are a range of opinions in that discussion, including those of a professional MD PhD who uses internet surveys to collect data for her peer-reviewed research articles.

I can understand that there is reluctance to use internet surveys because there is potential for misuse. On the other hand, this is true of any scientific method - all techniques will give erroneous results if improperly used. What is most important is that your particular survey is well designed and appropriate to provide the data needed to address your scientific question, and that you anticipate and address the concerns with your survey (Is there any reason to believe that anyone answered your survey more than once? If people responding are from different countries, would that affect your results? etc.)

In fact, internet surveys are widely used and accepted in research. Hundreds of peer-reviewed papers using internet surveys have been published and many more are published every month. I pasted some examples below. There are many, many more.

For the statistics, I would recommend starting simple (comparing averages & standard deviations) and then moving to more complex tests for further analysis. It's tough to recommend specific tests without knowing more about your data and question, so please post back with additional info if you need more help. You may want to direct statistics-related questions to the math forum.

Good luck!
Chris

Keeling ML, Piercy FP
A careful balance: Multinational perspectives on culture, gender, and power in marriage and family therapy practice
JOURNAL OF MARITAL AND FAMILY THERAPY 33 (4): 443-463 OCT 2007

Nikken P, Jansz J, Schouwstra S
Parents' interest in videogame ratings and content descriptors in relation to game mediation
EUROPEAN JOURNAL OF COMMUNICATION 22 (3): 315-336 SEP 2007

Kypri K, Paschall MJ, Maclennan B, et al.
Intoxication by drinking location: A web-based diary study in a New Zealand university community
ADDICTIVE BEHAVIORS 32 (11): 2586-2596 NOV 2007

Smith SY, Curbow B, Stillman FA
Harm perception of nicotine products in college freshmen
NICOTINE & TOBACCO RESEARCH 9 (9): 977-982 SEP 2007

Adams T
Producers, directors, and horizontal communication in television news production
JOURNAL OF BROADCASTING & ELECTRONIC MEDIA 51 (2): 337-354 JUN 2007

Togari T, Yamazaki Y, Nakayama K, et al.
Development of a short version of the sense of coherence scale for population survey
JOURNAL OF EPIDEMIOLOGY AND COMMUNITY HEALTH 61 (10): Art. No. 921 OCT 2007

Im EO, Chee W, Tsai HM, et al.
Internet communities for recruitment of cancer patients into an internet survey: A discussion paper
INTERNATIONAL JOURNAL OF NURSING STUDIES 44 (7): 1261-1269 SEP 2007

Kinney WC, Benninger MS
Assessment of quality of life among patients with sinonasal disease as determined by an Internet survey based on the Rhinosinusitis Disability Index
ENT-EAR NOSE & THROAT JOURNAL 86 (8): 482+ AUG 2007

Ducak AS, Mandic ML, Blazeka M
Internet survey on basic eating and living habits in adults
ANNALS OF NUTRITION AND METABOLISM 51: 131-131 P107 Suppl. 1 2007

Wasson JH, MacKenzie TA, Hall M
Patients use an internet technology to report when things go wrong
QUALITY & SAFETY IN HEALTH CARE 16 (3): 213-215 JUN 2007
Times Cited: 0
Context Sensitive Links View full text from group of Free Journals hosted by Highwire

Bordes M, Semjen F, Sautereau A, et al.
Which anaesthesia for children undergoing MRI? An internet survey in the French university hospitals
ANNALES FRANCAISES D ANESTHESIE ET DE REANIMATION 26 (4): 287-291 APR 2007

Hill EL, Cumming RG, Lewis R, et al.
Sleep disturbances and falls in older people
JOURNALS OF GERONTOLOGY SERIES A-BIOLOGICAL SCIENCES AND MEDICAL SCIENCES 62 (1): 62-66 JAN 2007
Times Cited: 1
Context Sensitive Links

Cohrs JC, Maes J, Moschner B, et al.
Determinants of human rights attitudes and behavior: A comparison and integration of psychological perspectives
POLITICAL PSYCHOLOGY 28 (4): 441-469 AUG 2007

Andrews L, Kiel G, Drennan J, et al.
Gendered perceptions of experiential value in using web-based retail channels
EUROPEAN JOURNAL OF MARKETING 41 (5-6): 640-658 2007

Potoglou D, Kanaroglou PS
Household demand and willingness to pay for clean vehicles
TRANSPORTATION RESEARCH PART D-TRANSPORT AND ENVIRONMENT 12 (4): 264-274 JUN 2007

Damschroder LJ, Roberts TR, Zikmund-Fisher BJ, et al.
Why people refuse to make tradeoffs in person tradeoff elicitations: A matter of perspective?
MEDICAL DECISION MAKING 27 (3): 266-280 MAY-JUN 2007

Givens JL, Houston TK, Van Voorhees BW, et al.
Ethnicity and preferences for depression treatment
GENERAL HOSPITAL PSYCHIATRY 29 (3): 182-191 MAY-JUN 2007

Smith DM, Loewenstein G, Rozin P, et al.
Sensitivity to disgust, stigma, and adjustment to life with a colostomy
JOURNAL OF RESEARCH IN PERSONALITY 41 (4): 787-803 AUG 2007

It is true that peer-reviewed journals publish studies using internet based surveys. It is true that you can do a statistically valid survey on the internet. However, to do so requires much plannning and careful design. In order to make the results survey meaningful, certain steps must be taken. Take a look at the papers Chris posted and look for how they deal with the problems presented by doing an internet-based study. Hopefully you can find some techniques to apply to your study to make it more meaningful.