Content analysis may be defined as a methodology by which the researcher seeks to determine the manifest content of written, spoken, or published communications by systematic, objective, and quantitative analysis. It is a means of trying to learn something about people by examining what they write, produce on television, or make movies about.
It is used to investigate the content of the Mass Media (although it has applications across a wider range of spheres – the analysis of historical documents, for example). In other words, it’s used to explore the content of various media (books, magazines, TV, film etc.,) in order to discover how particular issues are presented.
At its most basic, content analysis is a statistical exercise that involves categorising some aspect or quality of people’s behaviour and counting the number of times such behaviour appears (a simple content analysis might involve counting the number of minutes men and women appear on screen in a programme such as “Big Brother”). In this way, content analysis helps us to build-up a picture of the patterns of behaviour that the social interaction portrayed in the media.
SAMPLING
Sampling techniques for content analysis throw up a set of specific issues in addtition to those found in other quantitative methods. These are influenced by two factors:
1. Topic
2. Unit of analysis (the material)
CENSUS
If you are studying an event or a topic that has a discrete time period and a manageable amount of material, it may be possible to take a census – that is to look at all of the occurrences of this material. One example might be particular billboard advertising campaign that might have just a few different examples.
EVENT-DRIVEN
You may be studying the coverage of a particular event that occurs over a specific time period, for example, a political election. There is need to consider the time-frame of the event in order to decide the sample. Clearly, there will be more coverage as the election approaches and at certain peak times. There is a need to think carefully about the hypothesis in order to construct a sample that addresses the relevant aspects.
To continue with, if you are looking at the coverage of a particular political organisation over a year, you will need to take any such peaks into account. You might need to think about how you would analyse the coverage of the key issues and then construct a sample period in which relatively little is happening as a comparison.
SAMPLING CONSIDERATIONS: PRINT
If the topic involves the analysis of newspapers or magazines there are certain features that will need to be aware of in order to construct a good sample.
ANALYSIS CONSIDERATIONS: PRINT
As we have mentioned, deciding on you unit of analysis will depend very much on your hypotheses. If you are looking at newspapers, some of the things you may want to consider are whether you treat an entire article as a unit or whether you break this down. For example: You may consider only the headline and the first leader paragraph to be of interest, or you may break the article down into separate paragraphs.
You may also consider whether you are simply interested in articles that appear on the front page or whether the whole newspaper is of interest, bearing in mind that this will also contain in-depth articles, opinion pieces, editorials and letters.
You may also want to give some thought to the use of images that accompany a piece of text, particularly in a newspaper article.
If you are looking at the representation of a phenomenon in a fictional genre, will you need to consider all of the relevant texts or can you break this down to make a representative sample. For example: By analysing the top ten best sellers.
SAMPLE CONSIDERATIONS: TELEVISION
Broadcasting also has a set of considerations when considering sampling issues. Some of which are broadly analogous to those for print media.
1. Editorial stance: Some television networks have particular idiosyncrasies in their programming and editorial style. The nature and depth of news coverage may depend on the ethos of the editors. But another consideration may be whether the station is publicly funded or commercial.
2. Target audience: Particular networks or particular programmes will appeal to a particular audience. News coverage in particular will vary according to time of day and length of programme. Programming and content will also vary according to the nature of the network.
3. Broadcast time: Time of day will clearly have implications for programming. Particular times may be assumed mainly to be aimed at a specific audience demographic – people who are working from home, children, insomniacs – and this will affect not only the types of programme broadcast but also the content of programmes that appear regularly throughout any given day, particularly news broadcasts.
4. Reach: Again, related to target audience and broadly analogous regionally, while others are available nationally or internationally. This will also be influenced by whether they are transmitted terrestrially or via satellite or cable.
ANALYSE CONSIDERATIONS: BROADCAST
Analysing broadcast material can be very complex given the interplay of script and visual material. If you are looking at a news broadcast for example, you will need to think about how to break this into units of analysis. Where does one segment end and another being? In a recent project looking at the coverage of the 2003 war in Iraq, the coders spent many merry hours agonising over the differences between a one minute piece to camera by a journalist and a ten second introduction by the studio anchor.
The imagery in a piece of television obviously provides a great deal of the context and careful thought is needed to take this into account.
PROCEDURE
The process of conducting a content analysis essentially involves designing a questionnaire, except that, instead of people answering the question, you will get the answers from the text that you are analysing.
This is known as a coding sheet and the answers can be very simple or very complicated, depending on the research questions and hypothesis.
Each unit of analysis will need a separate coding sheet. You can either spend time in a darkened room surrounded by piles of paper ticking off each attribute of the text, or you can construct an electronic version of your coding sheet and enter data straight to the computer for subsequent analysis.
CONSTRUCTING A CODING SHEET
As in a questionnaire, you can divide a coding sheet into discrete sections.
The first is analogous to the demographics section. This will analyse elements of the medium itself that you feel will make a difference to the ways in which the issue is covered. Some things you might want to look at are:
1. When and where- year, date or day of the week; time of broadcast. Does coverage change over time? What publication or television outlet the piece appeared in.
2. Prominence- word length or timing. Position in the publication or in the broadcast.
3. Context – what other material surrounds the piece you are interested in? For example, what are the themes in the other stories on a front page, or what are the preceding and following items in a news broadcast?
The second major section will directly concern your research questions/hypotheses. This will need extremely careful planning.
Again, some things will be simple counts, but others may involve the identification of themes and language that will need to be operationally defined to avoid straying into the realms of more qualitative analysis.
1. Simple counts:
Speakers – who gets to speak directly?
Sources – who is quoted?
Themes – what are the key themes or messages running through the text?
2. Linguistic modifiers:
Labels – are there recurring labels used for a particular group or issue?
Language of the headline – are there any recurring words?
3. Imagery:
What are the most common images that accompany that issue?
PILOTING
· Content analysis can be vey time-consuming so it is important that you get it right. It’s worth taking a small sample of text and testing your coding sheet before you begin the entire analysis.
· Check that your coding sheet works and that you haven’t missed anything important. As a tip, it can be useful to have ‘other’ section so that you can fill in details of occurrences that were unexpected and then add this to the final version of the coding sheet.
· Remember that all categories must be mutually exclusive, i.e. something that is coded can only be coded in one way. Often there will be a blurring of the categories so you should take that time to decide which category an example might fit into now.
· It is worth experimenting at this stage to some extent.
· The final version of the coding sheet is one that you must stick to. You must follow the rules of quantitative methods – this is not an exploratory method, and you cannot change the coding sheet as you go through the final analysis.
INTER-CODER RELIABILITY
· A final, and essential, check on your coding sheet is that another person would code a text in the same way that you have. This will avoid accusations of subjectivity, but will also help your construct the final version so that categories are clear and mutually exclusive.
· Many pieces of published content analysis work will include a mathematical check on inter-coder reliability that quantifies the amount of agreement between different coders. As part of an introduction to this technique it’s not strictly necessary, but is really is worth taking a selection of examples of our text and get at least one other person to code them.
· Get together with your other coders and compare the notes.
ANALYSIS
· Since this is a purely descriptive method, the good news is that the analysis you wil use will be descriptive statistics.
· Frequencies and cross-tabulation will be sufficient to describe your finding and can prove very powerful.
STRENGTHS
1. It is relatively easy to gain access to the broadcast or publication you want to study.
2. It is relatively easy and inexpensive to build a representative sample.
3. It produces highly reliable (usually quantitative) data. Content Analyses are usually easy to repeat (“replicate”). Complex forms of social interaction can be quantified using a standardised framework (the content analysis grid) that can be applied across a wide range of media.
4. It can present an objective account of events, themes, issues and so forth that may not be immediately apparent to a reader, viewer or general consumer.
5. It is an unobtrusive method - it doesn’t involve the researcher interacting with the people / things being studied. The researcher cannot, therefore, influence the behaviour of the people being studied.
LIMITATIONS
1. May not be as objective as it claims since the researcher must select and record data accurately. In some instances (such as a television programme) the researcher must make choices about how to interpret particular forms of behaviour (for example, when a character is acting aggressively). For example, the researcher decides things like: what categories will be used and whether or not everyone be put neatly into a particular category).
2. By attempting to quantify behaviour (such as the relationship between people), this method may not tell us very much about the quality of people’s relationships.
3. May be time-consuming (for example, analysing a range of newspapers or TV programmes in detail).
4. As with all statistical data, it provides us with a snapshot of people’s behaviour at a single moment in their life.
5. It describes, rather than explains, people’s behaviour. It does not tell us what behaviour means to those involved and those watching.
CONCLUSION
The development of qualitative data analysis programs for personal computers has flourished in the past decade. Specialized programs for various levels of coding, theory building, annotating and displaying data are available.
Content analysis has many applications in research. However, there are several challengers because of many options and no straight forward guidelines. As noted earlier, content analysis project is unique and requires tailored procedures to fully explore the richness of the data. In addition, the procedure may be extremely time-consuming even for those well versed in its procedures.
When software is used to assist in the categorization and organization of data, the researcher must evaluate every segment of text for contest and subtleties not detectable by computer. It has been suggest that novices should contact experienced content analysts who can supply support, tips new ideas, and suggest tricks of the trade quickly and easily. This is especially true when learning to use qualitative data analysis and software since no single existing program does everything that a researcher might want or need. Overcoming the challenges inherent in methodologies is well worth the effort because of their potential usefulness in research.
Source: Internet
It is used to investigate the content of the Mass Media (although it has applications across a wider range of spheres – the analysis of historical documents, for example). In other words, it’s used to explore the content of various media (books, magazines, TV, film etc.,) in order to discover how particular issues are presented.
At its most basic, content analysis is a statistical exercise that involves categorising some aspect or quality of people’s behaviour and counting the number of times such behaviour appears (a simple content analysis might involve counting the number of minutes men and women appear on screen in a programme such as “Big Brother”). In this way, content analysis helps us to build-up a picture of the patterns of behaviour that the social interaction portrayed in the media.
SAMPLING
Sampling techniques for content analysis throw up a set of specific issues in addtition to those found in other quantitative methods. These are influenced by two factors:
1. Topic
2. Unit of analysis (the material)
CENSUS
If you are studying an event or a topic that has a discrete time period and a manageable amount of material, it may be possible to take a census – that is to look at all of the occurrences of this material. One example might be particular billboard advertising campaign that might have just a few different examples.
EVENT-DRIVEN
You may be studying the coverage of a particular event that occurs over a specific time period, for example, a political election. There is need to consider the time-frame of the event in order to decide the sample. Clearly, there will be more coverage as the election approaches and at certain peak times. There is a need to think carefully about the hypothesis in order to construct a sample that addresses the relevant aspects.
To continue with, if you are looking at the coverage of a particular political organisation over a year, you will need to take any such peaks into account. You might need to think about how you would analyse the coverage of the key issues and then construct a sample period in which relatively little is happening as a comparison.
SAMPLING CONSIDERATIONS: PRINT
If the topic involves the analysis of newspapers or magazines there are certain features that will need to be aware of in order to construct a good sample.
- Editorial stance: Many publications will have a particular political or social stance that may affect the way that they portray certain issues.
- Target audience: The intended readership will also affect this coverage, not just in terms of appealing to particular beliefs or attitudes, but also in the type of coverage that an issue may receive. For example, if you are studying the reputation of a particular multinational company, there will be publications that feature that company in terms of ‘headline-grabbing’ stories or events and there will be others that place more emphasis on financial performance.
- Publication dates: The regularity with which the text is published will also affect content. You could reasonable expect a daily newspaper to be slightly different from one that is published at weekends only; perhaps in terms of the depth of converge. Daily newspapers will have their own consideration. If you are looking at sporting coverage for example, you may need to consider the differences daily. There may well be differences in the amount and type of coverage during the build-up to the event compared to the post-event reaction and analysis.
- Circulation: The reach of the publication will also be important, and this will include the three factors above. You could reasonably expect a regional publication to vary in editorial stance, target audience, frequency of publication and depth of coverage of particular issues from those of a national or international publication.
ANALYSIS CONSIDERATIONS: PRINT
As we have mentioned, deciding on you unit of analysis will depend very much on your hypotheses. If you are looking at newspapers, some of the things you may want to consider are whether you treat an entire article as a unit or whether you break this down. For example: You may consider only the headline and the first leader paragraph to be of interest, or you may break the article down into separate paragraphs.
You may also consider whether you are simply interested in articles that appear on the front page or whether the whole newspaper is of interest, bearing in mind that this will also contain in-depth articles, opinion pieces, editorials and letters.
You may also want to give some thought to the use of images that accompany a piece of text, particularly in a newspaper article.
If you are looking at the representation of a phenomenon in a fictional genre, will you need to consider all of the relevant texts or can you break this down to make a representative sample. For example: By analysing the top ten best sellers.
SAMPLE CONSIDERATIONS: TELEVISION
Broadcasting also has a set of considerations when considering sampling issues. Some of which are broadly analogous to those for print media.
1. Editorial stance: Some television networks have particular idiosyncrasies in their programming and editorial style. The nature and depth of news coverage may depend on the ethos of the editors. But another consideration may be whether the station is publicly funded or commercial.
2. Target audience: Particular networks or particular programmes will appeal to a particular audience. News coverage in particular will vary according to time of day and length of programme. Programming and content will also vary according to the nature of the network.
3. Broadcast time: Time of day will clearly have implications for programming. Particular times may be assumed mainly to be aimed at a specific audience demographic – people who are working from home, children, insomniacs – and this will affect not only the types of programme broadcast but also the content of programmes that appear regularly throughout any given day, particularly news broadcasts.
4. Reach: Again, related to target audience and broadly analogous regionally, while others are available nationally or internationally. This will also be influenced by whether they are transmitted terrestrially or via satellite or cable.
ANALYSE CONSIDERATIONS: BROADCAST
Analysing broadcast material can be very complex given the interplay of script and visual material. If you are looking at a news broadcast for example, you will need to think about how to break this into units of analysis. Where does one segment end and another being? In a recent project looking at the coverage of the 2003 war in Iraq, the coders spent many merry hours agonising over the differences between a one minute piece to camera by a journalist and a ten second introduction by the studio anchor.
The imagery in a piece of television obviously provides a great deal of the context and careful thought is needed to take this into account.
PROCEDURE
The process of conducting a content analysis essentially involves designing a questionnaire, except that, instead of people answering the question, you will get the answers from the text that you are analysing.
This is known as a coding sheet and the answers can be very simple or very complicated, depending on the research questions and hypothesis.
Each unit of analysis will need a separate coding sheet. You can either spend time in a darkened room surrounded by piles of paper ticking off each attribute of the text, or you can construct an electronic version of your coding sheet and enter data straight to the computer for subsequent analysis.
CONSTRUCTING A CODING SHEET
As in a questionnaire, you can divide a coding sheet into discrete sections.
The first is analogous to the demographics section. This will analyse elements of the medium itself that you feel will make a difference to the ways in which the issue is covered. Some things you might want to look at are:
1. When and where- year, date or day of the week; time of broadcast. Does coverage change over time? What publication or television outlet the piece appeared in.
2. Prominence- word length or timing. Position in the publication or in the broadcast.
3. Context – what other material surrounds the piece you are interested in? For example, what are the themes in the other stories on a front page, or what are the preceding and following items in a news broadcast?
The second major section will directly concern your research questions/hypotheses. This will need extremely careful planning.
Again, some things will be simple counts, but others may involve the identification of themes and language that will need to be operationally defined to avoid straying into the realms of more qualitative analysis.
1. Simple counts:
Speakers – who gets to speak directly?
Sources – who is quoted?
Themes – what are the key themes or messages running through the text?
2. Linguistic modifiers:
Labels – are there recurring labels used for a particular group or issue?
Language of the headline – are there any recurring words?
3. Imagery:
What are the most common images that accompany that issue?
PILOTING
· Content analysis can be vey time-consuming so it is important that you get it right. It’s worth taking a small sample of text and testing your coding sheet before you begin the entire analysis.
· Check that your coding sheet works and that you haven’t missed anything important. As a tip, it can be useful to have ‘other’ section so that you can fill in details of occurrences that were unexpected and then add this to the final version of the coding sheet.
· Remember that all categories must be mutually exclusive, i.e. something that is coded can only be coded in one way. Often there will be a blurring of the categories so you should take that time to decide which category an example might fit into now.
· It is worth experimenting at this stage to some extent.
· The final version of the coding sheet is one that you must stick to. You must follow the rules of quantitative methods – this is not an exploratory method, and you cannot change the coding sheet as you go through the final analysis.
INTER-CODER RELIABILITY
· A final, and essential, check on your coding sheet is that another person would code a text in the same way that you have. This will avoid accusations of subjectivity, but will also help your construct the final version so that categories are clear and mutually exclusive.
· Many pieces of published content analysis work will include a mathematical check on inter-coder reliability that quantifies the amount of agreement between different coders. As part of an introduction to this technique it’s not strictly necessary, but is really is worth taking a selection of examples of our text and get at least one other person to code them.
· Get together with your other coders and compare the notes.
ANALYSIS
· Since this is a purely descriptive method, the good news is that the analysis you wil use will be descriptive statistics.
· Frequencies and cross-tabulation will be sufficient to describe your finding and can prove very powerful.
STRENGTHS
1. It is relatively easy to gain access to the broadcast or publication you want to study.
2. It is relatively easy and inexpensive to build a representative sample.
3. It produces highly reliable (usually quantitative) data. Content Analyses are usually easy to repeat (“replicate”). Complex forms of social interaction can be quantified using a standardised framework (the content analysis grid) that can be applied across a wide range of media.
4. It can present an objective account of events, themes, issues and so forth that may not be immediately apparent to a reader, viewer or general consumer.
5. It is an unobtrusive method - it doesn’t involve the researcher interacting with the people / things being studied. The researcher cannot, therefore, influence the behaviour of the people being studied.
LIMITATIONS
1. May not be as objective as it claims since the researcher must select and record data accurately. In some instances (such as a television programme) the researcher must make choices about how to interpret particular forms of behaviour (for example, when a character is acting aggressively). For example, the researcher decides things like: what categories will be used and whether or not everyone be put neatly into a particular category).
2. By attempting to quantify behaviour (such as the relationship between people), this method may not tell us very much about the quality of people’s relationships.
3. May be time-consuming (for example, analysing a range of newspapers or TV programmes in detail).
4. As with all statistical data, it provides us with a snapshot of people’s behaviour at a single moment in their life.
5. It describes, rather than explains, people’s behaviour. It does not tell us what behaviour means to those involved and those watching.
CONCLUSION
The development of qualitative data analysis programs for personal computers has flourished in the past decade. Specialized programs for various levels of coding, theory building, annotating and displaying data are available.
Content analysis has many applications in research. However, there are several challengers because of many options and no straight forward guidelines. As noted earlier, content analysis project is unique and requires tailored procedures to fully explore the richness of the data. In addition, the procedure may be extremely time-consuming even for those well versed in its procedures.
When software is used to assist in the categorization and organization of data, the researcher must evaluate every segment of text for contest and subtleties not detectable by computer. It has been suggest that novices should contact experienced content analysts who can supply support, tips new ideas, and suggest tricks of the trade quickly and easily. This is especially true when learning to use qualitative data analysis and software since no single existing program does everything that a researcher might want or need. Overcoming the challenges inherent in methodologies is well worth the effort because of their potential usefulness in research.
Source: Internet
No comments:
Post a Comment
I will be glad, if you recall someday that you heard this news from this blog. Do spare some time to leave a comment.