Buy, sell, or hold? A sense-making account of factors influencing trading decisions

Figures & data

Figure 1. A typical task window from Experiment 1.

Note: The figure shows the non-conflicting condition with bad news and a negative trend.

Table 1. Results of Experiment 1

			Trend
			Positive	Negative
Panel A : Western group, N = 30
Trading latency	News valence	Good	48.33 (48.89)	60.00 (61.84)
		Bad	45.67 (38.28)	36.67 (24.75)
Share number	News valence	Good	1.35 (0.92)	0.83 (0.96)
		Bad	1.03 (1.01)	0.4 (0.81)
Returns	News valence	Good	3.14 (2.00)	−3.31 (2.22)
		Bad	2.12 (1.38)	−2.68 (1.16)
Panel B: Eastern group, N = 30
Trading latency	News valence	Good	92.33 (71.98)	106.34 (70.78)
		Bad	66.00 (55.27)	73.66 (61.45)
Share number	News valence	Good	1.13 (0.96)	1.15 (0.917)
		Bad	0.72 (0.96)	0.60 (0.87)
Returns	News valence	Good	4.24 (2.25)	−4.35 (2.25)
		Bad	2.94 (1.89)	−3.44 (2.02)

Notes: Panel A shows the results of the Western participant group, and Panel B shows the results of the Eastern participant group in Experiment 1. Data presented consists of trading latency (the number of data points that participants saw before they made a buy/sell decision, or the maximal number of points), Share number (the average number of shares participants had after making their decisions), and returns (the difference between the asset price at decision time and at the time of initial presentation of the series). The results are presented for each of the four experimental conditions separately.

Table 2. Hypotheses and results of Experiment 1

Hypothesis	Variables in statistical analysis		Results and conclusions
	Dependent variable	Independent variable
H1 : When provided with both types of information, people base their trading strategy on news more than they do on price graphs	Final Share number	Culture, trend, news valence, instance	The effect of news valence on final share number (F (1, 29) = 29.35; p < 0.001; partial η^2 = 0.50) was larger than that of the trend (F (1, 29) = 7.56; p = 0.01; partial η^2 = 0.21)
			We accepted H1
H2: People track prices more and show more active trading (buying or selling rather than holding their assets) in non-contradicting conditions than in contradicting ones	Deviation of final share number from 1, price tracking	Condition type (contradicting or non-contradicting)	NS
			We rejected H2
H3: People sell more assets when the news is bad than they buy when it is good when short-selling is not possible	The difference between final share number and 1	News valence	When news was good people bought fewer shares than they sold when news was bad (F (1,479) = 5.16; p = 0.024)
			We accepted H3
H4: Trading latency is shorter when uncertainty is higher, that is, when there is an inconsistency between news valence and price trends	Trading latencies	Condition type	NS
			We rejected H4
H5: The effect of news on trading latencies is stronger than that of the price trend, and trading latency is shorter when news is bad	Trading latency	Culture, trend, news valence, instance	Trading latency was longer when news was good (F (1, 29) = 29.05; p < 0.01; partial η^2 = 0.50) but the effect of trend was insignificant
			We accepted H5
H6: Participants from Eastern Asian countries react to news more than participants from Western countries in inconsistent conditions	Number of shares, trading latency and returns	Culture, condition type	NS
			We rejected H6
H7a: People from Eastern culture exhibit longer trading latencies	Trading latency	Culture, trend, news valence, instance	Trading latency was shorter for Western participants (F (1, 29) = 17.23; p < 0.01; partial η^2 = 0.37)
			We accepted H7a
H7b: Trading decisions of people from eastern countries result in higher degrees of dispersion in their returns	Return dispersion	Culture	Return dispersion in the Eastern group was larger than that of Western group
			(t (239) = 5.60; p < 0.001))We accepted H7b
H8 : People who are more open to experience have shorter trading latencies	Trading latencies	Self rating of openness to experience	Participants with greater openness to experience had lower trading latencies
			We accepted H8

Notes: This table summarizes the hypotheses examined in Experiment 1 and their results. NS represents insignificant results. The variable condition type had two values: contradiction condition or non-contradicting condition.

Figure 2. A typical task window from Experiment 2.

Note: The figure shows the conflicting condition with bad news and a positive trend.

Table 3. Results of Experiment 2

Western participants (N = 30)			Trend
			Positive	Negative
Trading latency	News valence	Good	75.00 (58.87)	69.00 (60.78)
		Bad	67.33 (55.11)	45.00 (38.90)
Share number	News valence	Good	1.15 (0.97)	0.85 (0.95)
		Bad	0.55 (0.87)	0.31 (0.72)
Plausibility	News valence	Good	0.67 (0.16)	0.65 (0.17)
		Bad	0.65 (0.18)	0.68 (0.17)
Forecasts	News valence	Good	0.76 (0.66)	0.69 (1.05)
		Bad	−0.08 (0.94)	−0.49 (1.07)
Returns	News valence	Good	2.20 (1.29)	−2.23 (1.36)
		Bad	1.57 (1.23)	−1.59 (0.91)

Notes: This table shows averages of the results of Experiment 2. Data presented consists of trading latency (the number of data points that participants saw before they made a buy/sell decision, or the maximal number of points), Share number (the average number of shares participants had after making their decisions), Plausibility ratings (participants’ rating of the plausibility that presented news could affect the price of each asset), forecasts (the difference between participants’ forecasts and the last data point), and returns (the difference between the asset price at decision time and at the time of initial presentation of the series). The results are presented for each of the four experimental conditions separately.

Table 4. Hypotheses and results of Experiment 2

Hypothesis	Variables in statistical analysis		Results and conclusions
	Dependent variable	Independent variable
H1 : When provided with both types of information, people base their trading strategy on news more than they do on price graphs	Final Share number	Trend, news valence, instance	The effect of news on final share number (F (1, 29) = 11.47; p = 0.002; partial η^2 = 0.28) was larger than the effect of trend (F (1, 29) = 4.54; p = 0.042; partial η^2 = 0.14)
			We accepted H1
H2: People track prices more and show more active trading (buying or selling rather than holding their assets) in non-contradicting conditions than in contradicting ones	Deviation of final share number from 1, final number of shares	Condition type (contradicting or non-contradicting)	NS
			We rejected H2
H3: People sell more assets when the news is bad than they buy when it is good when short-selling is not possible	The difference between final share number and 1	News valence	Participants sold more shares when news was bad and the trend in the price data negative than they bought when news was good and the trend in the price data was positive (F (1, 239) = 24.20; p < 0.001)
			We accepted H3.
H4: Trading latency is shorter when uncertainty is higher, that is, when there is an inconsistency between news valence and price trends	Trading latencies	Condition type	NS
			We rejected H4
H5: The effect of news on trading latencies is stronger than that of the price trend, and trading latency is shorter when news is bad	Trading latency	Trend, news valence, instance	Trading latency was longer when the news was good (F (1, 29) = 8.23; p = 0.008; partial η^2 = 0.22). No main effect of trend was obtained
			We accepted H5
H9: There is a positive correlation between views about the extent to which an event would affect prices and final share number	Plausibility estimates	Trend, news valence, instance	NS
			We rejected H9
H10: The difference between a participant’s forecast and the last data point depends on the news valence and the direction of the trend in the price data	The difference between forecasts and asset prices	Trend, news valence, instance	The difference between forecasts and asset prices was higher when news was good (F (1, 29) = 38.93; p < 0.001; partial η^2 = 0.57), and when the trend was positive (F (1, 29) = 14.76; p = 0.001; partial η^2 = 0.34)
			We accepted H10
H11: There is a positive correlation between the difference between a participant’s forecast and the last data point, and the final share number	Final share number	The difference between a participant’s forecast and the last data point	There were positive correlations when the trends were positive, whether the news items were good (r = 0.53; p < 0.001) or bad (r = 0.48; p < 0.001)
			We accepted H11

Notes: This table summarizes the hypotheses examined in Experiment 2 and their results. NS represents insignificant results. The variable condition type had two values: contradiction condition or non-contradicting condition.