Reliability and validity of a standard survey I had eleven people play five versions of a video game in single sessions, for a total of 55 sessions (11x5). I used the Task Evaluation Questionnaire from the Intrinsic Motivation Inventory to capture their experiences.
Do I need to test the reliability and validity of this survey for the data that I collected? Someone said that I may have to and I don't know how it's done, or whether these 55 different sessions are enough. 
On the website the authors say:

The IMI items have often been modified slightly to fit specific activities. Thus, for example, an item such as "I tried very hard to do well at this activity" can be changed to "I tried very hard to do well on these puzzles" or "...in learning this material" without effecting its reliability or validity.

This makes me think that I don't have to, but I don't know much about survey research. What are your thoughts? 
Here's a follow-up question:
Even if you think that I don't need to, let's say I really want to compute reliability and validity of this survey for my experiment. How would I do that? Would the data that I already have (55 sessions) be enough? I'm asking because collecting new data is very inconvenient at the moment. 
Thank you so much.
 A: Regarding your first question (computing reliability estimates on your own data vs. rely on published reliabilities), several methodologists recommend using your own data, in particular because reliability can vary depending on the population (see a reference on this below). In a way, reliability is a property of a particular set of scores, not of a scale per se. In the long term, reporting reliability estimates is also useful for things like reliability generalization studies.
Vacha-Haase, T., Kogan, L.R., & Thompson, B. (2000). Sample compositions and variabilities in published studies versus those in test manuals: Validity of score reliability inductions. Educational and Psychological Measurement, 60 (4), 509-522. 
A: Reliability and validity are both multifaceted concepts, so be careful about using them in such general terms.
Your friend is probably asking about the internal consistency (reliability) of the survey. This is commonly measured using Cronbach's alpha. However, Cronbach's alpha is typically used on single-factor surveys; the TEQ has four factors. You could compute four reliability coefficients, one for each factor, or a confirmatory factor analysis might be more appropriate.
Truth be told, I would consider most of this overkill. The beauty of using a standardized test is that its reliability has been confirmed many times before, hence standardization. That is, of course, assuming the TEQ is a commonly used standardized test-- I haven't used it myself. If this were a write-up, citing an authoritative source in regards to the reliability of the TEQ would probably be sufficient to appease most readers. If not, calculating Cronbach's alpha for each factor and saying something like "all alpha > .7" should be sufficient (provided of course, all alpha are greater than .7).
The validity of the survey depends on what you're doing with the data. The test is valid to the extent that it measures what you're trying to measure, or predicts what you're trying to predict. As long as the connection between the test and your use of it is plausible, you probably don't need to compute any statistics for validity.
