SEARCH

SEARCH BY CITATION

Keywords:

  • Internet;
  • search engine;
  • suicidal ideation;
  • suicide;
  • World Wide Web

Abstract

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Cross-correlation was examined for the volume of suicide-related Internet searches and suicide death rate. Analysis of Google data and figures released by the Ministry of Health, Labour, and Welfare indicated that the volume of searches using the search terms jisatsu (suicide) and jisatsu houhou (suicide method) are not correlated with the suicide death rate. In addition, a rising suicide death rate might be related to the increase in suicide-related search activity (particularly utsu[depression]), but an increase in suicide-related search activity itself is not directly linked to the rise of suicide death rate.

WITH THE WIDESPREAD use of the Internet, more than three-quarters of people in Japan have come to use this technology in their daily lives. While an earlier meta-analysis highlighted the influence of the media on suicide,1 the Internet has caused cluster suicides in a manner identical to that by other media such as television and newspapers. In 2008, suicide using hydrogen sulfide – a method introduced over the Internet – involved more than 1000 people. In the early 2000s, the so-called Internet suicide pacts, which are essentially group suicides committed by people who become acquainted with each other over the Internet, occurred frequently. Thus, penetration of the Internet into society might increase cluster suicide.

In contrast, some researchers have suggested that the Internet provides resources for suicide prevention and support.2,3 For example, self-help group activities mediated through Web communities are made available by people who have contemplated suicide.2 Further, other efforts are being made to provide psychoeducation over the Internet to people who have attempted suicide and people who have entertained suicidal thoughts.3

Therefore, the Internet can be effective in both encouraging and preventing suicides. Regarding the suicide-related use of the Internet, the use of search engines can be an indicator of the present situation. Hitherto, the results of several studies have indicated that the volume of searches using suicide-related search terms is positively correlated with the suicide death rate.4,5 None of those studies, however, have clarified the causal relationship between search volume and suicide death rate. It is possible to interpret the situation in either of two ways: either searches using suicide-related terms influence the suicide death rate, or the increasing number of suicide deaths raises people's interest in suicides, which leads to an increase in search volumes. Therefore, in order to examine these two opposing hypotheses, the present study examined the cross-correlation coefficient between the volumes of searches involving suicide-related search terms and the suicide death rate.

METHOD

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

The study examined the data within a time range of 6 years between 2004 and 2009. This was determined on the basis of the period for which search volume data have been provided by Google Insights for Search (GIS) in Japan.

Data on search volumes using suicide-related terms were collected using GIS. GIS is a service that provides data on search volumes based on words and phrases. It analyzes the search terms entered by people into the Google search engine worldwide. This service is equipped with means to ensure the validity of its data, such as showing only the results of searches exceeding a certain volume and excluding searches made by a single user within a short period. GIS can also display search volume data graphically within a specified time range and region, with conversion so that the volume is displayed on a scale of 0–100, the highest search volume being displayed as 100 (Fig. 1). Data were collected on the relative monthly volume of searches using the search terms jisatsu (suicide), utsu (depression), and jisatsu houhou (suicide method) in Japan from January 2004 to December 2009. These search terms were selected with regard to a preceding study on suicide-related search terms4,5 and their search volumes.

image

Figure 1. Google Insights for Search result screen. Vertical axis indicates how many searches have been done for a particular term. Users can download the data to a CSV file by clicking the button at the top of the results page. This graphical content was retrieved from http://www.google.com/insights/search/?hl=en-US#q=suicide&geo=JP&date=1%2F2004%2061m&cmpt=q

Download figure to PowerPoint

The monthly suicide death rate was then applied to the demographic statistics released by the Ministry of Health, Labour, and Welfare.

The data were analyzed by calculating the cross-correlation coefficient between the monthly suicide death rate and the volume of searches using the suicide-related search terms. The time lag range (3 months before to 3 months after) was determined on the basis of the preceding studies on the media's influence on suicide.6

RESULTS

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Analysis (Table 1) indicated that only the search term ‘utsu’ (depression) was significantly correlated with the suicide death rate. The search terms ‘jisatsu’ (suicide) and ‘jisatsu houhou’ (suicide method) did not show a significant correlation with the suicide death rate. The volume of searches using the search term ‘utsu’ (depression) was positively correlated with the suicide death rate in the same or previous month and was negatively correlated with the suicide death rate after 3 months.

Table 1.  Completed suicide and suicide-related search terms in Japan (2004–2009)
Time lag‘Suicide’ (Jisatsu)‘Depression’ (Utsu)‘Suicide method’ (Jisatsu houhou)
  • *

    P < 0.05;

  • **

    P < 0.01

  • Cross-correlation was calculated between search volumes and from 3 months before to 3 months after suicide.

−30.090.31**−0.12
−2−0.030.43**−0.05
−10.040.42**0.09
00.070.25*0.09
10.010.090.06
2−0.11−0.18−0.05
3−0.19−0.30**−0.13

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

The present results suggest that the volume of searches with the terms ‘jisatsu’ (suicide) and ‘jisatsu houhou’ (suicide method) is not correlated with the suicide death rate. Although there are new suicide risks due to the dissemination of suicide methods over the Internet, such as cluster suicide using briquette coal, they are unlikely to develop to the point of influencing the suicide death rate at a statistical level. Considering that the volume of searches involving the search term ‘utsu’ (depression) is positively correlated, to some extent, with the suicide death rate in the same or previous month, it is reasonable to suppose that rising suicide death rate might be related to the increase in suicide-related search activity, but increase in suicide-related search activity itself is not directly linked to the rise of suicide death rate.

Further, the volume of searches using the search term ‘utsu’ (depression) was negatively correlated with the suicide death rate after 3 months. This suggests that the increase in the search volume using the term ‘utsu’ (depression) leads to a declining suicide death rate after 3 months. Currently, suicide prevention efforts in Japan concentrate on handling depression, and many of the high-ranking search results and key word-targeted advertisements displayed for the search term ‘utsu’ (depression) introduce various websites providing psycho-education for handling depression and resources for support. Given the negative correlation between the search volumes of ‘utsu’ (depression) and the suicide death rate after 3 months, such websites could have effectively prevented suicides.

Finally, some limitations of the present study must be mentioned. First, the present study examined the issue over a limited time range and was based on searches in a single language; therefore, generalization of the results requires some caution. Second, previous studies indicated that internet usage-related suicides exercise a greater negative influence over young people with low media literacy.7 Further, given that the data provided by GIS could not be divided according to user age, the difference in the relationship between suicide death rate and search term volumes by age was not examined in this research. Future research is necessary to resolve this technological problem in order to estimate the relationship between Internet searches and suicide death rate with more precision.

Although the present study had certain limitations, the results suggest that suicide-related Internet search activities are unlikely to increase suicide death rates. The results contradict the threatening image of the Internet generated by Internet suicide pacts and the hydrogen sulfide gas suicides, which were executed using Internet information.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Stack S. Media impacts on suicide: A quantitative review of 293 findings. Soc. Sci. Q. 2000; 81: 957971.
  • 2
    Eichenberg C. Internet message boards for suicidal people: A typology of users. Cyberpsychol. Behav. 2008; 11: 107113.
  • 3
    Sueki H. User characteristics of suicide-preventive information service on the internet and an effect of browsing the contents on users' suicidal ideation: Comparison between minors and adults. Suicide Prev. Crisis Interv. 2010; 30: 2330 (in Japanese).
  • 4
    McCarthy MJ. Internet monitoring of suicide risk in the population. J. Affect. Disord. 2010; 122: 277279.
  • 5
    Sueki H. Relationship between completed suicide and search volumes of suicide-related terms on the Internet: An investigation using Google Insights for Search beta. Jpn. J. Clin. Psychol. 2011; 11: 7782 (in Japanese).
  • 6
    Ishii K. Measuring mutual causation. Soc. Sci. Res. 1991; 20: 188195.
  • 7
    Thompson S. The Internet and its potential influence on suicide. Psychiatr. Bull. 1999; 23: 449451.