Research Methods in IS: Hypotheses

Research Methods in IS: Research Hypothesis(es)

• A hypothesis is a tentative, intelligent guess to guide problem-solving
• Hypotheses can originate from sub-problems, with a one-to-one correspondence
• The number of hypotheses equals the number of sub-problems
• Hypotheses can be directional (suggesting a positive or negative relationship) or non-directional (no direction stated)
• Researchers can only state directional hypotheses when prior information (from experience, literature review, or existing theories) on the phenomena under study exists
• Exploratory studies may omit hypotheses

Research Methods in IS: Research Hypothesis(es) - Continued

• Hypotheses are neither proven nor disproven, but rather supported or not supported by data
• Trying to prove a hypothesis biases research, as the researcher might only look for confirming data
• Hypothesis acceptance or rejection depends solely on the data
• Incorrect educated guesses about the investigation outcome can be discovered if the data does not support the hypothesis

Research Methods in IS: Research Hypothesis(es) - Continued

• Examples of testable hypotheses:
  • Students receiving counseling show greater creativity increases than those not receiving counseling
  • Automobile A performs as well as Automobile B
• Hypotheses should be capable of objective verification and testing
• A hypothesis outlines what the researcher is looking for and is a testable proposition to assess its validity

Research Methods in IS: Characteristics of Hypothesis(es)

• Hypotheses should be clear and precise, for reliable inferences
• Hypotheses should be testable, allowing the derivation of confirmable or disconfirmable deductions through observation
• Hypotheses should specify the relationship between variables (if it's a relational hypothesis)

Research Methods in IS: Characteristics of Hypothesis(es) - Continued

• Hypotheses should be limited in scope, since narrower hypotheses are usually more testable
• Hypotheses must explain the facts underlying the need for explanation, utilizing related knowledge and accepted generalizations to deduce the initial problem
• The hypothesis needs empirical support, explaining what it claims

Research Methods in IS: Characteristics of Hypothesis(es) - Continued

• Hypotheses should be stated in simpler terms for easy understanding
• Consistency with known facts is necessary (based on a substantial body of established facts)
• The hypothesis should be readily testable within a reasonable time frame

Research Methods in IS: Null and Alternative Hypotheses

• A null hypothesis proposes no difference or relationship between groups or variables, essential in statistical analysis
• A null hypothesis is the opposite of the hypothesis being researched and is often tested for rejection
• Example: Comparing Method A to Method B and stating they perform equally well exemplifies a null hypothesis

Research Methods in IS: Null and Alternative Hypotheses - Continued

• Example research hypothesis: Teenagers in Program A graduate from high school at a higher rate than those in Program B
• The null hypothesis is "No difference in high school graduation rates between Program A and Program B"

Research Methods in IS: Null and Alternative Hypotheses - Continued

• Social workers trying to disprove a null hypothesis, in favor of their research hypothesis, is illustrated using the example of high school graduation rates
• This leads to acceptance or rejection of the research hypothesis based on the supporting data

Research Methods in IS: Null and Alternative Hypotheses - Continued

• An alternative hypothesis represents the result if the null hypothesis is not supported, essentially the set of possibilities beyond the null hypothesis, and it is often the hypothesis that one wishes to support or prove

Research Methods in IS: Null and Alternative Hypotheses - Continued

• Choosing the null and alternative hypotheses occurs prior to sample data collection
• Researchers should avoid creating a hypothesis based on their collected data and then testing the hypothesis on the same data

Research Methods in IS: Null and Alternative Hypotheses - Conclusion

• The alternative hypothesis is the one that researchers usually wish to prove, whereas the null hypothesis is the one researchers aim to disprove
• An example of the use of hypotheses: O.J Simpson murder trial
• One set of jurors believed the null hypothesis (O.J. innocent) based on collected evidence
• A second set of jurors rejected the null hypothesis and reached a verdict of guilty

Research Methods in IS: Errors in Hypothesis Testing

• Errors are possible when testing a hypothesis
• Type I error (false positive): Rejecting a true null hypothesis
• Type II error (false negative): Not rejecting a false null hypothesis

Research Methods in IS: Types of Errors

• Type I error is rejecting a true null hypothesis
• Type II error is accepting a false null hypothesis

Research Methods in IS: Type I Error

• Type I error, also known as a "false positive," is the mistake of rejecting a true null hypothesis.
• It involves accepting an alternative hypothesis when the results are due to chance
• Occurs when observation shows a difference when it does not exist statistically

Research Methods in IS: Type II Error

• Type II error, also called a "false negative," is that of not rejecting a false null hypothesis.
• Involves rejecting an alternative hypothesis when there is adequate power to accept it
• Occurs when failing to observe a difference when a difference truly exists