Questioning the Class Data: • Are the measurements valid? - What would make them NOT valid? • Are the measurements reliable? - What would make them NOT reliable? • Is there any bias in your data? - What would create bias in this activity?
1. Validity • A variable is a VALID measure of a property if it is relevant or appropriate as a representation of that property. • Sometimes it is better to use a rate than a count. - Percentages of population versus number of people in population. • Example: Your study would not be valid if you measured the height of students in an attempt to study students’ math grades.
2. Predictive Validity • How accurate does one variable predict another? • Does a variable predict success to a given task? • Example: Scores from IQ Tests are used to predict intelligence.
3. Errors in Measurement ? True Weight Scale stuck this morning and read a pound lower Measured Value Scale Reading True Value Random Error EXAMPLE: Bathroom Scale Scale always reads 3 pounds higher = = + + Bias + +
4. Types of Errors • Bias: When a measurement process systematically overstates or understates the true value. - Example: When measuring everyone’s height, all rulers started at 1 inch, not 0 inches. • Reliability: When there is a small random error after repeated measurements. - A study is reliable if measurements are consistent! - You want small variability (very little spread for your data).
5. Improving Reliability (reducing error) • Find the most accurate method of measuring • Repeat Measurements (obtain a larger sample) - If possible, take the averages of measurement data to draw conclusions from!
6. Class Example • You take your blood pressure at home using a home monitor. You get the following results: 120/80, 132/90, 125/85, 110/70, 135/85. You go to the doctor and find that your actual blood pressure is 121/80. -Is your blood pressure monitor a valid method for measuring your blood pressure? Explain. - Does your BP monitor have a problem with reliability or bias? Explain.