+ Reply to Thread
Results 1 to 7 of 7

Thread: difference between ANOVA and T-test

  1. #1
    Points: 3,999, Level: 40
    Level completed: 25%, Points required for next Level: 151

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    difference between ANOVA and T-test




    Hello,

    my teacher asked me last week if you can use ANOVA to investigate the relation between two variables instead of the unpaired T-test which you normally use ... I said you can choose between these two because they lead to the same result. Is this true?

  2. #2
    Points: 9,022, Level: 63
    Level completed: 91%, Points required for next Level: 28

    Posts
    38
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Quote Originally Posted by Leslie View Post
    Hello,

    my teacher asked me last week if you can use ANOVA to investigate the relation between two variables instead of the unpaired T-test which you normally use ... I said you can choose between these two because they lead to the same result. Is this true?
    Yes.

    "One factor analysis of variance (Snedecor and Cochran, 1989) is a special case of analysis of variance (ANOVA), for one factor of interest, and a generalization of the two-sample t-test. The two-sample t-test is used to decide whether two groups (levels) of a factor have the same mean. One-way analysis of variance generalizes this to levels where k, the number of levels, is greater than or equal to 2."

    http://www.itl.nist.gov/div898/handb...on3/eda354.htm

  3. #3
    Points: 3,999, Level: 40
    Level completed: 25%, Points required for next Level: 151

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Thanks!

    But why does 'Statistica' use these two options instead of only ANOVA?
    If the T-test gives us the same results as ANOVA, isn't it crazy? If you always use ANOVA, the T-test becomes unnecessary or am I seeing this the wrong way?

    Greets

  4. #4
    Points: 9,022, Level: 63
    Level completed: 91%, Points required for next Level: 28

    Posts
    38
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Quote Originally Posted by Leslie View Post
    Thanks!

    But why does 'Statistica' use these two options instead of only ANOVA?
    If the T-test gives us the same results as ANOVA, isn't it crazy? If you always use ANOVA, the T-test becomes unnecessary or am I seeing this the wrong way?

    Greets
    As far as I know, all the information you would get from a t-test can be gotten from ANOVA, _assuming_ that the two samples have equal variances. I don't know if ANOVA can be tweaked to give the same info a unequal-variances t-test would give.

    It could also just be that the Statistica t-test gives both 2-sided, and both 1-sided results, and ANOVA just gives the 2-sided without doing the calculations to provide the 1-sided results as well. Have not used Statistica, I don't know.

  5. #5
    TS Contributor
    Points: 17,636, Level: 84
    Level completed: 58%, Points required for next Level: 214
    JohnM's Avatar
    Posts
    1,948
    Thanks
    0
    Thanked 6 Times in 5 Posts
    Although t and ANOVA can give the same results, the t-test gives you the ability to do one- or two-tailed tests.

  6. #6
    Points: 1,682, Level: 23
    Level completed: 82%, Points required for next Level: 18

    Posts
    16
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: difference between ANOVA and T-test

    I'm sorry to bump this old topic, but I assumed it would be prefered to making a new topic about this subject.

    Is the result of an independent t-test and a ANOVA on 2 treatments ALWAYS the same? Even when you have a small sample size?

    Reason I'm asking this because I'm not sure if I fully understand this piece I found:

    ''First, a little history about this curious name. William Gosset (1876-1937) was a Guinness Brewery employee who needed a distribution that could be used with small samples. Since the Irish brewery did not allow publication of research results, he published under the pseudonym of Student. We know that large samples approach a normal distribution. What Gosset showed was that small samples taken from an essentially normal population have a distribution characterized by the sample size. The population does not have to be exactly normal, only unimodal and basically symmetric. This is often characterized as heap-shaped or mound shaped.''

    From:
    'For n > 30, the differences are negligible'
    http://askville.amazon.com/sample-si...estId=52854836

    Does this mean when you have n<30 the difference between a normal distribution and a t-distribution are not neglible? I'm pretty sure I'm missing something because don't the independent t-test and ANOVA both assume a normal distribution?

    But the main thing I need to know if a independent t-test and ANOVA always give the same result. Googling around says it is the case (t squared = F).

    Thanks in advance

  7. #7
    Devorador de queso
    Points: 95,705, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,931
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: difference between ANOVA and T-test


    Quote Originally Posted by BBD View Post
    Does this mean when you have n<30 the difference between a normal distribution and a t-distribution are not neglible? I'm pretty sure I'm missing something because don't the independent t-test and ANOVA both assume a normal distribution?
    Yes they both assume normality. But that doesn't mean that the sampling distribution of your statistics will be perfectly normal because we don't have a perfect estimate of the variance. That's why we use the t distribution. Also if you have a large enough sample size then you don't really need to worry about the assumption of normality (depending on what your goals are).

    But the main thing I need to know if a independent t-test and ANOVA always give the same result. Googling around says it is the case (t squared = F).
    Yes they will give the same results (assuming you use the same assumptions for both). Also note that t^2_{df=n} = F_{1,n}. In other words a squared t with n degrees of freedom is equal to an F with 1 numerator degrees of freedom and n denominator degrees of freedom. If you think about the standard one way anova situation the numerator degrees of freedom represent the number of groups you're comparing minus one. So if we're doing a t-test we're comparing 2 groups and this seems to work out nicely.

+ Reply to Thread

           




Similar Threads

  1. Replies: 3
    Last Post: 09-16-2010, 06:27 AM
  2. Replies: 2
    Last Post: 07-26-2010, 05:37 PM
  3. Replies: 1
    Last Post: 06-26-2010, 07:55 AM
  4. ANOVA, Tukey and finding the difference
    By Tystat in forum Other Software
    Replies: 0
    Last Post: 05-11-2010, 05:53 AM
  5. Replies: 1
    Last Post: 07-24-2008, 08:27 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats