Validity and Reliability of the

¶ … validity and reliability of the research outcomes. In order to provide the best possible research results, the researcher therefore must make decisions regarding the appropriateness of any particular research design to the study in question. This creates quite a challenge for the researcher, as both research design and research questions are fluid in nature. Neither is a predetermined fit for the other.

With a question such as "Do negative ads run by candidates depress voter turnout?" It is therefore important to determine the research design that can most accurately and truthfully answer the question. Because this question determines the general attitudes of voters, it was decided that the experimental research design could yield some accurate results. This design is significantly flexible in determining the sample of participants, as well as the question being studied.

If such research were to be designed, there are a number of factors that need to be taken into account to ensure a valid causal relationship and to minimize the likelihood of a spurious relationship. One significant factor is the attitude of voters towards politicians. Voters may feel negative towards all politicians, especially in the light of recent criticism against them and their ability to indeed effect the changes they had promised. This, more than negative advertising, could be a significant factor in lower voter turnout.

The experimental design is well-equipped to handle this type of eventuality by means of the control group and the time between survey sessions. A group of randomly selected voters may for example receive a questionnaire to determine their attitude to politics, politicians and voting before being exposed to negative advertising. The group can then be redivided to create a relatively similar experimental and control group, where the former is subsequently exposed to negative advertising for a number of sessions and the latter is not.

It is also possible that voters could change their attitudes after having completed the questionnaire. This could affect the validity of the results for the experimental group. To handle this eventuality, several questionnaires can be delivered over time for the whole group before the experimental group is selected. A mean can then be determined to most accurately divide the group into a similar experimental and control group. After the experimental group is determined, negative advertisements can be displayed.

In order increase the validity of this experiment, a number of survey questionnaires over time can also be delivered in order to minimize non-visible influences such as an extreme reaction to a particular person in a particular advertisement. At the same time, the same number of questionnaires can be delivered to the control group. A mean of all the questionnaires over the determined time can then provide relatively accurate data regarding whether voters are indeed influenced to the extent of changing the decision to vote particularly by negative advertising.