I've struggled with this over the years myself. At its extreme, of course culture impacts IQ scores. That means that what we think of as intelligence here in the US means a lot in some places and rather little in others. But you aren't likely to be taking the WAIS or WISC kit to another country. Instead you'll test people in the US. Some have recently come from another country, and some have lived here a while but still move between ethnic/"old world" settings and mainstream/"new world" settings.

You can not simply dismiss IQ scores you don't like, as I've seen people do at times, as invalid due to cultural differences. You should be able to cite specific cultural issues, traditions, or conflicts with the client you tested, and cite evidence to document them, before you dismiss test scores. I had a woman from Canada I tested once who scored poorly on Vocabulary. Thinking the results for that subtest likely were invalid, and to make her feel better, I said, "The words at the end were pretty hard, especially if you grew up speaking French." She grew up in Canada in an area where everyone spoke English, so the score was probably accurate and I probably made her feel worse.

• 1) English fluency
  This is one of those, "Well, of course!" points, but it means more than giving a child who speaks Spanish at home a Spanish WISC. It means realizing that their Spanish is contaminated with English, and language fluency resources are split between two languages. It is possible that neither will be as well-developed as in a native speaker of one language.
  
  Also relevant is "language deprivation," or the amount and kind of verbal stimulation at home. This is tied to parenting style, poverty, single versus dual parent homes, number and age of siblings, family emotional styles, availability of extended family… middle class parents talk more, with longer sentence, introducing more constructs and vocabulary, as well as more complex sentence structures.
  
  Language also means your ability to communicate with the client, using words that match their vocabulary when you question, and allowing them to answer with words they are comfortable with. This strengthens the rapport you build and the validity of the assessment.

  
  

• 2) Orientation to time, and time and performance pressure
  Some children are highly motivated to "make the cutoff" and finish within a time limit, and others are not. Other would not question an authority figure or an adult if they did not understand or agree with the task. While not always due to cultural variables, these kinds of differences can be.

  
  

• 3) Meaningfulness and familiarity with the task
  Remember Ceci's children and the circle-square-triangle task versus the butterfly-bird-bee game? Familiarity with the content is important, as is teaching the task and making sure the child understands the instructions. Did you let the child know that guessing was OK, speed was important, and re-prompt when needed?

Think too about the content of items. When emotional experiences, expressions, and customs differ, how are anger and joy alike, a friend and an enemy?

• 4) Motivation, persistence, and conscientiousness in performance
  A number of ideas come into play here; expectation of success, effort optimism, parent training of achievement behavior, view of school as a place associated with success or failure, perceived threat of prejudicial judgment, expectation that the results are truly assessments of one's ability… all come into play. Talking about a "problem-solving test," giving reinforcement for effort, asking about school experiences during the interview, and about real-world coping skills is all important.

  
  

• 5) Individualistic versus collective issues
  This gets to what is "your" work versus what is "our work." Native American children are often noted to show more awareness of this in their interactions with adults. You might also see this in someone who's answer is, "I'd ask my brother." She or he may be used to answering what they know, and seeking help immediately on what they don't as it is faster and more efficient. Getting the child to understand you want to know what they think or know is sometimes harder. Family emphasis placed on academic achievement and individual success is also a key here.

  
  

• 6) Beliefs about there being "one correct answer" to problems
  This can be a cultural thing, but also an aspect of how the interaction is viewed. "That's right, you do eat them both (and apple and a banana) and they are also fruits" may not mean "Change your thinking style" to some people.
  
  Sometimes this is linked directly to quality of school experiences, enrichment in the home environment, and opportunities in the community. You should consider environmental variables as well as internal ones to explain test findings.

  
  

• 7)Acculturation
  Mercer (1971) classified IQ scores for Mexican and African American children by level of acculturation. Between the highest and lowest levels of acculturation, Mexican American children showed a 20 point difference (104.4 and 84.5). African American children showed a 17 point difference (99.5 and 82.7).

Others have noted the interaction between examiner and client. Older and Older, Older and Younger, Younger and Older, and Younger and Younger can create different expectations in different cultures, just like the four matches of male and female. These spell out what is appropriate in the setting. Thus, sometimes a client might not correct your misunderstanding of them so as not to be rude, but other times may not say anything because they perceive you as rude.

• 8) Test Norms
  Examine the subject sample for representativeness in both the design, construct validation, norming of the test (remember the High School Personality Questionnaire?), and predictive validity studies.