LIS 397'1 Introduction to Research in Library and Information Science Summer, 2003 Thoughtful Thursd

1
LIS 397.1Introduction to Research in Library and
Information ScienceSummer, 2003Thoughtful
Thursday Day 14
2
Limitations of t tests

• Can compare only two samples at a time
• Only one IV at a time (with two levels)
• But you say, Why dont I just run a bunch of t
  tests?
• Its a pain in the butt.
• You multiply your chances of making a Type I
  error.

3
ANOVA

• Analysis of variance, or ANOVA, or F tests, were
  designed to overcome these shortcomings of the t
  test.
• An ANOVA with ONE IV with only two levels is the
  same as a t test.

4
ANOVA (contd.)

• Remember back to when we first busted out some
  scary formulas, and we calculated the standard
  deviation.
• We subtracted the mean from each score, to get a
  feel for how spread out a distribution was how
  DEVIANT each score was from the mean. How
  VARIABLE the distribution was.
• Then we realized if we added up all these
  deviation scores, they necessarily added up to
  zero.
• So we had two choices we coulda taken the
  absolute value, or we coulda squared em. And we
  squared em.
• S(X M)2

5
ANOVA (contd.)

• S(X M)2
• This is called the Sum of the Squares (SS). And
  when we add em all up and average them (well
  divide by N-1), we get S2 (the variance).
• We take the square root of that and we have S
  (the standard deviation).

6
ANOVA (contd.)

• Lets work through the Hinton example on p. 111.

7
F is . . .

• F is the variance ratio.
• F is
• between conditions variance/error variance
• (systematic differences error variance) /error
  variance
• Between conditions variance/within conditions
  variance
• (This from Hinton, p. 112, p. 119.)

8
Check out . . .

• ANOVA summary table on p. 120. This is for a ONE
  FACTOR anova (i.e., one IV). (Maybe MANY
  levels.)
• Sample ANOVA summary table on p. 124.
• Dont worry about unequal sample sizes
  interpretation of the summary table is the same.
• The only thing you need to realize in Chapter 13
  is that for repeated measures ANOVA, we also
  tease out the between subjects variation from the
  error variance. (See p. 146 and 150.)
• Note, in Chapter 15, that as factors (IVs)
  increase, the comparisons (the number of F
  ratios) multiply. See p. 167, 174.
• Memorize the table on p. 177. (No, Im only
  kidding.)

9
Interaction effects

• Heres what I want you to understand about
  interaction effects
• Theyre WHY we run studies with multiple IVs.
• A significant interaction effect means different
  levels of one IV have different influences on the
  other IV.
• You can have significant main effects and
  insignificant interactions, or vice versa (or
  both sig., or both not sig.) (See p. 157, 158.)

10
Chi square test

• Lets work example in Hinton, p. 243.
• Just know that you use the chi square test when
  you have FREQUENCY data.

11
Lets talk about the final
12
Some sample problems
13
Homework

• Continue reading.
• Final next Wednesday.
• See you Tuesday.